Source: The Conversation – UK – By Julian Hargreaves, Senior Lecturer, Department of Sociology and Criminology, City St George’s, University of London
A man believed to be Jihad Al-Shamie, a 35-year-old British citizen born in Syria, has been shot dead by police after launching an attack on a synagogue in Manchester on Yom Kippur, the holiest day in the Jewish calendar. Melvin Cravitz, 66, and Adrian Daulby, 55, died in the attack – one having been accidentally shot by police trying to stop the suspect.
According to BBC News, a member of the public called the police at 9:31am to report the incident. Greater Manchester Police deployed firearms officers to the scene at 9:34am. At 9:38am officers declared “Operation Plato” – a code word used by UK emergency services for a marauding terrorist attacker. At 9:39am, armed counter terrorism police officers, shot and killed Al-Shamie who died at the scene. Counter terrorism police later confirmed the attacked as a “terrorist incident”.
Within hours, it had become clear that many foresaw such an attack. The Financial Times reported comments from Marc Levy, chief executive of the Jewish Representative Council, a body representing Jewish communities in Greater Manchester. Levy described the events as “an inevitability”.
The Board of Deputies of British Jews, a national body representing Jewish communities across the UK, described the attack as “sadly something we feared was coming”.
The Jewish Chronicle, a Jewish interest newspaper, reported that staff at the London Centre for the Study of Contemporary Antisemitism were “shocked but not surprised”.
Recent research by the thinktank Antisemitism Policy Trust analysed demonstrations against the war in Gaza. It found public expressions of anti-Jewish hatred alongside more legitimate pro-Palestinian and anti-Israeli government sentiment, including Arabic chants referencing the massacre of Jews in 628BC.
The Community Security Trust, an organisation serving and protecting Jewish communities, records and reports antisemitic incidents in the UK. In 2023, the CST recorded 4,296 incidents – the largest number in a single year. CST used previous lower annual totals to explain how antisemitism is now fuelled by responses to the October 7 Hamas attacks: 1,684 incidents in 2020, 2,261 in 2021 and 1,662 in 2022.
The CST works carefully to investigate and verify all reports of antisemitism. While their work is entirely robust, it cannot easily reveal whether the dramatic rise in incidents reflects growing antisemitic sentiment, or increases in the reporting of antisemitic incidents to the CST, or both.
According to Home Office figures, religious hate crime against Jewish people more than doubled between the years ending March 2023 to March 2024. In 2022-23, there were 1,543 incidents recorded by the police. In 2023-24, there were 3,282.
While the number of incidents is lower than those against Muslim people – 3,432 in 2022-23 and 3,866 in 2023-24 – Jewish people are more likely to suffer religious hate crime. There were 121 incidents for every 10,000 Jewish in England and Wales compared to 10 incidents for every 10,000 Muslim people.
The same caveats apply here. We cannot know whether these increases represent growing hostility towards Jewish people in the UK or more Jewish people reporting hostility to the police. This issue is further complicated by the fact that police-recorded crime is no longer regarded as meeting the standard required of reliable national statistics due to poorly managed recording practices.
How widespread is antisemitism in the UK?
In 2017, the Institute for Jewish Policy Research (JPR) published what is arguably the most robust mapping of antisemitism in the UK. It estimated the extent of anti-Jewish attitudes using a nationally representative survey.
The JPR found that around 2% of the UK population might be labelled as “hardcore” antisemites and a further 3% as “softer” antisemites on the basis that both groups hold multiple antisemitic ideas. It also found that at least one more antisemitic idea is held by 30% of British society.
It is difficult to say with clarity whether or not antisemitism is rising in the UK, mainly because police statistics are so unreliable. But when terrorist attacks occur, we seek to understand what has happened and reach for robust information. This creates an urgent need for fresh research with better police data and more recent crime data.
Regardless of all this, findings from the JPR show that while strong antisemitism remains relatively uncommon in the UK, the odds of Jewish people encountering neighbours with at least one antisemitic idea remains worryingly high. Small wonder then that so many felt this attack was just a matter of time.
Julian Hargreaves does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
A SEPTA train moves along the Market-Frankford Line in West Philadelphia.AP Photo/Matt Rourke
On April 13, 1967, around 1:30 p.m., Lt. Joseph Larkin of the Philadelphia Police Department’s subway unit visited the Philadelphia High School for Girls to interview the school’s librarian, 61-year-old Miriam S. Axelrod.
Axelrod had written a letter to Mayor James H.J. Tate about poor conditions on Philadelphia’s Broad Street Line subway. In her letter, she stated that the escalators in the subway concourse of the Walnut-Locust station were out of operation for several weeks and requested that they “be put in running order.”
Axelrod also asked that “something be done” about people using the subway stairs “as a latrine.”
As a historian of post-1968 Philadelphia, a proud alumna of Girls’ High and a rider of Philadelphia’s mass transit, Southeastern Pennsylvania Transportation Authority – more commonly known as SEPTA – I was thrilled to find Axelrod’s story among 1960s administrative reports to the police commissioner in the city archives.
Axelrod’s story reminds us that for nearly a century, Philadelphia’s mass transit has been plagued by poor conditions and unstable funding. Commuters’ complaints have often convinced government officials to act. However, no effective plan has ever been implemented to definitively solve the city’s transit crises.
SEPTA’s current turmoil
On Sept. 15, 2025, SEPTA fully restored its service, by court order, after implementing 20% service reductions and a 21.5% fare increase due to a US$213 million budget deficit.
Passengers at Olney Transportation Center in North Philadelphia board a SEPTA bus on Aug. 25, 2025, a day after major service cuts went into effect. AP Photo/Matt Rourke
Some lawmakers have argued that SEPTA is guilty of mismanaging funds, since the agency already received over $1 billion in state subsidies last year for operating assistance and asset improvement.
Public transport in the 1920s
As a longtime Philadelphian who lived in Center City, Miriam Axelrod was familiar with the strengths and shortcomings of public transportation.
At the age of 4, her family emigrated from Russia to Carmel, New Jersey. By 1920, they made Philadelphia their home just as the city’s Russian community became the largest immigrant group due to Jewish people escaping pogroms in Europe. Axelrod grew up living in South and North Philadelphia.
At that time, dozens of private transit companies operated in Philly.
Southern Penn operated city buses. Red Arrow provided suburban trolley service. Pennsylvania and Reading railroads offered high-speed rail lines. The Philadelphia Rapid Transit Co. alone brokered deals with 64 underlying companies to annually rent their services under 999-year leases. Fiscal responsibility for quality transportation was complicated and often dependent on public funding.
During PRT’s early years, it paid the city $15,000 annually for snow removal. In return, the city spent $2 million for street paving and bridge repairs.
That was the same year Axelrod graduated from William Penn High School for Girls. Her classmates keenly noted in their yearbook that she had a “remarkable capacity for starting arguments” in which “any debatable subject will do.”
Six years later, the first segment of the Broad Street Subway traveling from Olney Station in North Philadelphia to City Hall opened to the public. Unlike the bus, trolley and railways systems, the city owned the El, short for elevated, line and subway. The city leased both systems to PRT and made the transit company responsible for their maintenance.
On Jan. 1, 1940, the Philadelphia Transportation Co., a private company with a 21-member board of directors that included five city representatives including the mayor, Robert E. Lamberton, merged the transit companies and took over PRT’s operations. PTC became responsible for 10,000 employees and providing transportation for 2 million passengers a day.
PTC also acquired extensive financial responsibilities. Payroll expenses cost $327,000 each week. The annual rate for leasing the subway and El was roughly $3 million. PTC had to provide its 25,000 bondholders an annual income of at least $959,207 while also fulfilling its promise to offer modern transit vehicles.
Overcrowding and frequent fare increases
During the 1940s through the 1960s, Axelrod took public transportation to her job as a librarian at Central High School and later Frankford High School.
Meanwhile, PTC made good on its promise to provide better transit service. In its first eight years of operation, PTC spent $22.8 million to purchase 1,506 new streetcars, buses and trackless trolleys while also improving terminal and plant facilities. The company even purchased advertisements in The Philadelphia Inquirer to highlight its achievements. PTC extended 38 existing routes and created 18 new routes that serviced old residential and industrial areas, along with newly developed neighborhoods.
By 1949, however, many of PTC’s 3.2 million daily riders were complaining about overcrowded subways, the end of free exchanges between popular routes and frequent fare increases.
Passengers ride a subway car in Philadelphia on Feb. 15, 1946. AP Photo
Both PTC and the city faced scrutiny for these issues, although each party had distinct transit obligations outlined in their joint contract. PTC had to provide “safe and adequate service” that included spending on maintenance and replacement of transit equipment. The city was responsible for police and fire services on mass transit along with auditing PTC’s records. Both parties had to agree on fare changes under the state Public Utility Commission’s supervision.
Nevertheless, when issues on mass transit occurred, the city could persuade PTC to improve conditions, but the city was only required to offer emergency services to commuters.
When Larkin personally addressed Axelrod’s 1967 complaint about the subway, he informed her that the United Elevator Co. was repairing the escalators. He also assured her that the subway unit arrested 45 to 50 intoxicated people each month because they were at risk of falling onto the subway tracks. In “isolated cases,” Larkin explained, police arrested people for public urination and defecation.
Larkin reassured Axelrod that PTC could keep subway conditions clean and under control. In reality, PTC was underwater in responsibilities and debt.
On Sept. 30, 1968, SEPTA, a state agency formed five years earlier, took over PTC and managed transportation for the city and its surrounding areas. SEPTA bought PTC for approximately $47.9 million, settling the company’s debt, accepting its pension liability and buying out the institution’s roughly 1.7 million shareholders. Now federal and state funding rather than fare revenue largely determined the quality of the city’s public transit.
Five counties in Greater Philadelphia contribute subsidies to SEPTA in exchange for transit service. Philadelphia alone contributes $110 million. State subsidies also help finance SEPTA’s $1.74 billion operating budget, while federal subsidies support SEPTA’s $1 billion capital budget to pay for major repairs and new equipment. State politicians annually vote on funding for SEPTA, but there has not been a concrete solution to the funding crisis.
However, Philadelphians never ceased to demand better transit service. During the 1980s, the Pennsylvania Public Interest Coalition established the Transit Riders Action Campaign, also known as TRAC, which advocated that SEPTA have better safety, funding, accountability, service and stable fares. The Transport Workers Union Local 234 advised TRAC, while several organizations partnered with them: the Action Alliance of Senior Citizens, the Clean Air Council, Disabled in Action and the Delaware Valley Interfaith Coalition.
Even today, local groups such as Save the Train with outspoken commuters – like Axelrod was in her day – have launched campaigns to halt service cutbacks and encourage residents to write and telephone legislators who can vote to fund SEPTA. Residents have consistently united to advocate for quality mass transit. All that remains is an agreement among lawmakers to make it possible.
Menika Dirkson does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
Peu d’idées dans la science moderne ont autant bouleversé notre compréhension de la réalité que l’espace-temps, cette trame entrelacée d’espace et de temps qui est au cœur de la théorie de la relativité d’Albert Einstein. L’espace-temps est souvent décrit comme le « tissu de la réalité ».
Le philosophe austro-britannique Ludwig Wittgenstein a un jour averti que les problèmes philosophiques surgissent lorsque « le langage part en vacances ». Il s’avère que la physique en est peut-être un bon exemple.
Au cours du siècle dernier, des mots usuels tels que « temps », « exister » et « intemporel » ont été réutilisés dans des contextes techniques sans que l’on examine le sens qu’ils ont dans le langage courant.
En philosophie de la physique, en particulier dans une approche connue sous le nom d’éternalisme, le mot « intemporel » est utilisé au sens littéral. L’éternalisme est l’idée que le temps ne s’écoule pas et ne passe pas, que tous les événements à travers le temps sont également réels dans une structure à quatre dimensions connue sous le nom d’« univers-bloc ».
L’éternalisme considère que tout existe de manière atemporelle et simultanée. (Rick Rothenberg/Unsplash), CC BY
Selon cette vision, toute l’histoire de l’univers est déjà écrite, de manière intemporelle, dans la structure de l’espace-temps. Dans ce contexte, « intemporel » signifie que l’univers lui-même ne perdure ni ne se déploie dans un sens réel. Il n’y a pas de devenir. Il n’y a pas de changement. Il n’y a qu’un bloc, et toute l’éternité existe de manière intemporelle à l’intérieur de celui-ci.
Mais cela conduit à un problème plus profond. Si tout ce qui se produit à travers l’éternité est également réel, et que tous les événements sont déjà là, que signifie réellement le fait que l’espace-temps existe ?
Un éléphant dans la pièce
Il existe une différence structurelle entre l’existence et l’occurrence. L’une est un mode d’être, l’autre, un mode d’arriver.
Imaginez qu’un éléphant se tienne à côté de vous. Vous diriez probablement : « Cet éléphant existe. » Vous pourriez le décrire comme un objet tridimensionnel, mais surtout, c’est un « objet tridimensionnel qui existe ».
En revanche, imaginez un éléphant purement tridimensionnel qui apparaît dans la pièce pendant un instant : un moment transversal dans la vie d’un éléphant existant, apparaissant et disparaissant comme un fantôme. Cet éléphant n’existe pas vraiment au sens ordinaire du terme. Il se produit. Il apparaît.
Un éléphant existant perdure dans le temps, et l’espace-temps catalogue chaque instant de son existence sous la forme d’une ligne mondiale en quatre dimensions – le parcours d’un objet dans l’espace et le temps tout au long de son existence. L’« éléphant qui apparaît » imaginaire n’est qu’une tranche spatiale du tube, un instant en trois dimensions.
Appliquons maintenant cette distinction à l’espace-temps lui-même. Que signifie l’existence d’un espace-temps à quatre dimensions au sens où l’éléphant existe ? L’espace-temps perdure-t-il dans le même sens ? L’espace-temps a-t-il son propre ensemble de moments « présents » ? Ou bien l’espace-temps – l’ensemble de tous les événements qui se produisent à travers l’éternité – est-il simplement quelque chose qui se produit ? L’espace-temps est-il simplement un cadre descriptif permettant de relier ces événements ?
Déjà des milliers d’abonnés à l’infolettre de La Conversation. Et vous ? Abonnez-vous gratuitement à notre infolettre pour mieux comprendre les grands enjeux contemporains.
L’éternalisme brouille cette distinction. Il traite toute l’éternité – c’est-à-dire tout l’espace-temps – comme une structure existante, et considère le passage du temps comme une illusion. Mais cette illusion est impossible si tout l’espace-temps se produit en un clin d’œil.
Pour retrouver l’illusion que le temps passe dans ce cadre, l’espace-temps à quatre dimensions doit exister d’une manière plus proche de l’éléphant à trois dimensions existant, dont l’existence est décrite par l’espace-temps à une dimension.
Chaque événement
Poussons cette réflexion un peu plus loin.
Si nous imaginons que chaque événement de l’histoire de l’univers « existe » dans l’univers-bloc, alors nous pourrions nous demander : quand le bloc lui-même existe-t-il ? S’il ne se déroule pas et ne change pas, existe-t-il hors du temps ? Si tel est le cas, alors nous ajoutons une autre dimension temporelle à quelque chose qui est censé être intemporel au sens littéral.
Pour donner un sens à cela, nous pourrions construire un cadre à cinq dimensions, en utilisant trois dimensions spatiales et deux dimensions temporelles. Le deuxième axe temporel nous permettrait de dire que l’espace-temps à quatre dimensions existe exactement de la même manière que nous considérons généralement qu’un éléphant dans une pièce existe dans les trois dimensions spatiales qui nous entourent, les événements que nous classons comme espace-temps à quatre dimensions.
À ce stade, nous sortons du cadre de la physique établie, qui décrit l’espace-temps à travers quatre dimensions seulement. Mais cela révèle un problème profond : nous n’avons aucun moyen cohérent de parler de ce que signifie l’existence de l’espace-temps sans réintroduire accidentellement le temps à travers une dimension supplémentaire qui ne fait pas partie de la physique.
C’est comme essayer de décrire une chanson qui existe à un moment donné, sans être jouée, entendue ou dévoilée.
De la physique à la fiction
Cette confusion façonne notre conception du temps dans la fiction et la science populaire.
Dans le film de James Cameron de 1984, The Terminator, tous les événements sont considérés comme fixes. Le voyage dans le temps est possible, mais la chronologie ne peut être modifiée. Tout existe déjà dans un état fixe et intemporel.
Dans le quatrième film de la franchise Avengers, Avengers : Endgame (2019), le voyage dans le temps permet aux personnages de modifier les événements passés et de remodeler la ligne temporelle, suggérant un univers en bloc qui existe et change à la fois.
Ce changement ne peut se produire que si la ligne temporelle à quatre dimensions existe de la même manière que notre monde à trois dimensions.
Mais indépendamment de la possibilité d’un tel changement, les deux scénarios supposent que le passé et l’avenir sont là et prêts à être parcourus. Cependant, aucun des deux ne s’intéresse à la nature de l’existence que cela implique, ni à la manière dont l’espace-temps diffère d’une carte des événements.
Comprendre la réalité
Lorsque les physiciens affirment que l’espace-temps « existe », ils travaillent souvent dans un cadre qui a discrètement brouillé la frontière entre existence et occurrence. Il en résulte un modèle métaphysique qui, au mieux, manque de clarté et, au pire, obscurcit la nature même de la réalité.
Rien de tout cela ne remet en cause la théorie mathématique de la relativité ou la science empirique qui la confirme. Les équations d’Einstein fonctionnent toujours. Mais la manière dont nous interprétons ces équations est importante, en particulier lorsqu’elle influence la façon dont nous parlons de la réalité et dont nous abordons les problèmes plus profonds de la physique.
Définir l’espace-temps est plus qu’un débat technique : il s’agit de déterminer dans quel type de monde nous pensons vivre.
Daryl Janzen ne travaille pas, ne conseille pas, ne possède pas de parts, ne reçoit pas de fonds d’une organisation qui pourrait tirer profit de cet article, et n’a déclaré aucune autre affiliation que son organisme de recherche.
Source: The Conversation – in French – By Fabrice Lollia, Docteur en sciences de l’information et de la communication, chercheur associé laboratoire DICEN Ile de France, Université Gustave Eiffel
Les élections africaines, déjà marquées par des tensions récurrentes autour de la transparence et de la désinformation, pourraient bientôt entrer dans une nouvelle ère, celle des deepfakes. Ces vidéos et audios générés par intelligence artificielle, capables d’imiter la voix, le visage et les gestes d’une personne avec un réalisme troublant, déplacent la frontière de la manipulation politique.
S’ils prêtent parfois à sourire lorsqu’ils mettent en scène des célébrités dans des détournements humoristiques, leur usage en période électorale représente une menace sérieuse pour la stabilité démocratique. Aux États-Unis, en Inde ou encore en Slovaquie, les deepfakes ont déjà été mobilisés pour influencer l’opinion publique. La question centrale est donc très simple : l’Afrique est-elle prête à affronter ce nouvel outil de manipulation électorale ?
Chercheur en sciences de l’information et de la communication, j’étudie la circulation de l’information, la désinformation et les vulnérabilités communicationnelles en contexte de crise. L’émergence des deepfakes illustre ces tensions. En Afrique, où la jeunesse hyperconnectée domine l’électorat mais où la culture numérique reste inégale, le risque est particulièrement élevé. J’en propose ici une lecture info-communicationnelle appliquée aux élections africaines
Des précédents inquiétants à l’échelle mondiale
Les deepfakes ne sont plus une hypothèse futuriste. Ils ont déjà marqué des épisodes électoraux clés, offrant des leçons pour les pays africains.
En 2023, en Slovaquie, quelques jours avant les législatives, un deepfake audio circulant sur Facebook et Telegram attribuait à Michal Simecka, chef du parti pro-occidental progressiste slovaque, une conversation où il planifiait de ruser le scrutin. Ce contenu a semé le doute au profit du camp populiste de Robert Fico. Il s’agit du premier cas documenté en Europe où un deepfake aurait pesé sur un scrutin national.
En 2024 aux États unis lors de la primaire démocrate du New Hampshire, des électeurs reçurent un appel téléphonique deepfake imitant la voix de Joe Biden et incident à l’abstention. Ce qui illustre l’usage des deepfakes pour dissuader la participation électorale, ce qui constitue une attaque frontale contre la démocratie.
En 2024 en Inde, les élections générales de 2024 ont été marquées par une explosion de deepfakes. Ces vidéos et sons générés par intelligence artificielle (IA) ont été diffusés massivement par les réseaux sociaux. Des acteurs de Bollywood ou même des personnalités politiques décédées ont été mis en scène pour soutenir ou attaquer les candidats.
Ces cas montrent que les deepfakes ne visent pas seulement à convaincre, mais surtout à introduire le doute, brouiller les repères et miner la confiance.
• Une faible culture de vérification : beaucoup d’utilisateurs partagent sans contrôler l’origine des contenus;
• Une viralité extrême : les messages et vidéos circulent rapidement dans les groupes fermés et sont difficiles à surveiller;
• Les institutions électorales sont contestées : la confiance citoyenne est fragile, ce qui confère une crédibilité accrue aux fausses informations.
Des signaux faibles apparaissent déjà :
Au Nigéria, en 2023, des inquiétudes ont émergé concernant la circulation de vidéos manipulées lors de la présidentielle.
Au Kenya, en 2022, TikTok et Facebook ont hébergé de nombreux contenus politiques manipulés, certains proches de techniques de falsification, dans le cadre de campagnes de désinformation.
L’Afrique se trouve donc dans une phase de vulnérabilité latente, réunissant tous les ingrédients pour que les deepfakes deviennent rapidement une arme politique.
À la différence des « fake news » classiques, les deepfakes tirent leur force de la synergie image/son créant une illusion sensorielle difficile à contester. Leur efficacité ne repose pas seulement sur la capacité à tromper, mais sur leur pouvoir de déstabilisation symbolique.
Ils peuvent ainsi créer un scandale contre un candidat, amplifier des clivages ethniques ou religieux et semer la confusion.
Cette érosion du contrat de vérité constitue une crise communicationnelle majeure qui fragilise les démocraties africaines déjà confrontées à des équilibres institutionnels précaires.
Une lecture info-communicationnelle
Les SIC permettent d’analyser ce phénomène sous un angle élargi. Trois axes sont particulièrement pertinents :
Tout d’abord, en termes de médiologie et de circulation des rumeurs, les deepfakes s’inscrivent dans une longue histoire des technologies de communication comme instruments de pouvoir. L’incertitude, le manque de transparence et l’opacité de certaines sphères d’information favorisent la prolifération de rumeurs, en particulier dans les contextes électoraux ou politiques.Les deepfakes ajoutent une couche technologique qui donne un vernis de crédibilité à la rumeur.
Ensuite, dans le cadre des logiques sociotechniques des plateformes, les algorithmes comme celui de TikTok privilégient les contenus sensationnels et polarisants. Dans ce système le deepfake devient une arme algorithmique amplifiée par l’économie de l’attention.
Enfin, on constate que dans un contexte africain marqué par des fractures linguistiques, éducatives et technologiques, la réception des deepfakes varie fortement. La culture numérique inégale favorise des appropriations différenciées, accentuant les asymétries de compréhension.
De nombreuses pistent émergent, mais leur mise en œuvre reste complexe :
Google, Meta ou Microsoft développent des outils capables d’identifier les contenus synthétiques. Mais ces technologies de détection restent coûteuses et rarement accessibles aux médias africains.
Des initiatives comme Africa Check jouent un rôle crucial en terme de médias et fact-checking, mais elles sont sous-dimensionnées face à la masse d’informations manipulées.
D’un point de vue juridique, certains pays africains légifèrent contre les fake news comme le Ghana ou l’Ouganda, mais il est à craindre que ces lois, dont l’encadrement est discutable, risquent de servir la censure politique plutôt que la protection citoyenne. Une approche panafricaine via l’Union Africaine ou les communautés régionales offrirait plus de crédibilité.
Former les jeunes et les moins jeunes à repérer, vérifier et questionner les contenus constitue un investissement démocratique stratégique. Les programmes scolaires universitaires, l’éducation aux médias, qui sont les leviers à long terme doivent intégrer la littérature numérique et médiatique comme compétences civiques.
Vers une souveraineté numérique africaine ?
La menace des deepfakes invite aussi à réfléchir à une souveraineté numérique africaine. L’Afrique ne peut pas dépendre uniquement des géants technologiques occidentaux pour sécuriser son espace informationnel. Le développement de laboratoires panafricains de recherche et de détection associés à des initiatives de société civile pourrait constituer une réponse endogène.
En outre, la coopération Sud-Sud ( par exemple entre l’Inde et certains pays africains) pourrait favoriser l’échange de solutions techniques et pédagogiques. Car il ne s’agit pas seulement de contrer la manipulation, mais aussi de bâtir une culture numérique partagée, capable de redonner confiance aux citoyens.
Les cas en Slovaquie, en Inde, aux États-Unis montrent que les deepfakes sont déjà une arme électorale redoutable. En Afrique, leur introduction dans le jeu politique n’est plus qu’une question de temps.
Mais la menace ne se réduit pas à une technologie. Elle révèle une vulnérabilité communicationnelle plus profonde qui se caractérise par une crise de confiance minant la légitimité démocratique. L’enjeu n’est donc pas uniquement de détecter les deepfakes, mais bien de reconstruire un rapport de vérité entre gouvernés et gouvernants.
Former les citoyens, renforcer les médias, développer une recherche locale et promouvoir une régulation panafricaine sont autant de pistes pour affronter ce défi. Car au-delà de la technique, c’est la capacité de l’Afrique à protéger l’intégrité de ses choix démocratiques qui est en jeu.
Fabrice Lollia does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
Source: The Conversation – Africa – By Cormac Price, Post-doctoral fellow the HerpHealth lab, office 218, Building G23. Unit for Environmental Sciences and Management, North-West University; University of KwaZulu-Natal
Black mambas (Dendroaspis polylepis) are Africa’s longest, most famous venomous snakes. Despite their fearsome reputation, these misunderstood snakes are vital players in their ecosystems. They keep rodent populations in check and, in turn, help to protect crops and limit disease spread. The species ranges widely across sub-Saharan Africa, from Senegal to Somalia and south into South Africa. They can adapt to many environments.
Zoologist Cormac Price, in new research with professors Marc Humphries and Graham Alexander and reptile conservationist Nick Evans, found that black mambas can be indicators of heavy metal pollution. We asked him about it.
How do black mambas indicate toxic pollution?
It’s about bio-accumulation. Bioaccumulation happens when chemicals, like pesticides or heavy metals, build up in an organism’s body. These toxins come from polluted environments, from waste products of human activities like manufacturing. They pollute water or soil and gradually accumulate in plants and animals.
If toxins are present in the environment, they may first be taken in by plants, and then by animals that eat the plants, and animals that eat those animals. Black mambas are quite high up the food chain, so a lot of the toxins would accumulate in their bodies. These poisonous substances can reach dangerous levels, causing health problems for whatever eats them.
We tested the presence of four types of heavy metals (arsenic, cadmium, lead and mercury) in the bodies of black mambas.
All our samples were from the eThekwini Municipality (greater Durban area) in South Africa. Durban is a busy shipping container port and has a large industrial sector that includes chemicals, petrochemicals and automotive manufacturing. Alongside all this industry the municipality also has a network of conservancies and green spaces, known as the Durban Metropolitan Open Space System.
We chose to test for these metals because they are widely used in different industries and can cause drastic negative effects in the body. Mercury primarily damages the nervous system, arsenic can cause cancer and skin lesions, cadmium harms kidneys and bones and lead mainly affects brain development and blood functions. Because these metals accumulate over time and are difficult to break down, even low-level exposure can lead to chronic poisoning and long-term health problems.
Black mambas appear to be doing well in Durban and taking advantage of the abundance of rodents, which they eat. Wherever there is human settlement there will be waste and discarded food which rodents take full advantage of. Black mambas can also be quite site-specific when not disturbed, living in the same refuge for many years, giving a clearer indication of pollution levels at that specific site. This makes the snakes potentially good bioindicator species.
A bioindicator species is one that helps us understand the health of an environment. Because they are sensitive to changes like pollution or habitat damage, their presence, absence or condition can reveal if an ecosystem is in good condition or is experiencing increases of pollution or degradation.
The pollutants can be detected and calculated from a non-invasive, harmless scale clipping. Snake scales are composed mostly of keratin, the same sort of protein that produces human hair and nails. To clip a very thin slice of snake scale is as harmless as clipping a human finger nail.
We collected 31 mambas that had already been killed by vehicles, people or dogs, and tested muscle and liver samples from them for toxins. We also took scale clippings from 61 live snakes.
This was the first time in Africa that a species of snake was tested to see if it could be used as an indicator species of heavy metal pollution.
What did you find?
We found that the heavy metal concentrations in scales correlated with those found in the muscle and liver samples. For three of the four metals, scales were as accurate for testing as muscle and liver samples. So the harmless testing method is as good as the more invasive one.
For arsenic, cadmium and lead, the snakes were accumulating significantly lower concentrations of these toxins in the open, natural sites of the Durban Metropolitan Open Space System compared to more industrial and commercial areas. Mercury was less significantly different due to its more volatile nature and its capacity to travel through the environment.
What made you test mamba scales in the first place?
In 2020, I attended a conference on amphibians and reptiles, where a friend of mine presented his work on heavy metal pollutants in tiger snakes in the city of Perth, Australia.
I’ve also been working with Nick Evans of KZN Amphibian & Reptile Conservation for some years, on urban reptile ecology. Nick began collecting scale clippings, and I began to realise, while looking through the literature, how novel this was on a continental scale. Snakes had never been tested as a potential bioindicator species of heavy metal pollution in Africa previously.
Marc Humphries is a professor of environmental chemistry, and I was aware of his work on lead exposure in Nile crocodiles at St Lucia, a wetland in South Africa. When he expressed interest in examining the scale clippings, we were thrilled. Graham Alexander’s expertise in snake behaviour in general and specifically snakes in Durban was also instrumental in the success of this research.
How can this help fight pollution?
The fight against pollution is in the hands of the municipality and city managers. What the snakes are doing is warning us of the increasing danger these pollutants pose to environmental health and ultimately human health. They are also showing us how important open spaces are to the overall environmental and human health of the city of Durban. The snakes are telling us a story; what people in authority decide to do with this story rests with them.
Nick Evans of KZN Amphibian & Reptile Conservation made valuable contributions to the research and was a co-author on the article.
Cormac Price does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
Scientists have always needed someone to help foot the bill for their work.
In the 19th century, for example, Charles Darwin made an expensive voyage to the southernmost tip of the Americas, visiting many other places en route, including his famous trek through the Galapagos Islands. The fossil evidence Darwin collected over his five-year journey eventually helped him to think about an infinite variety of species, both past and present.
The HMS Beagle and its crew traversed these places while testing clocks and drawing maps for the Royal Navy, and the voyage was funded by the British government. Darwin’s position as a naturalist aboard the ship was unpaid, but, fortunately, his family’s private assets were enough to cover his living expenses while he focused on his scientific work.
Today, government and private funding both remain important for scientific discoveries and translating knowledge into practical applications.
As a professor of science education, one of my goals while preparing future teachers is to introduce them to the characteristics of scientific knowledge and how it is developed. For decades, there has been a strong consensus in my field that educated citizens also need to know about the nature of the scientific enterprise. This includes understanding who pays for science, which can differ depending on the type of research, and why it matters.
Funding for science is more than just the amount of money. To a large extent, the organizations that fund research set the agenda, and different funders have different priorities. It can also be hard to see the downstream benefits of scientific research, but they typically outweigh the upfront costs.
Basic research leads to new knowledge
Basic research, also called fundamental research, involves systematic study aimed at acquiring new knowledge. Scientists often pursue research that falls into this category without specific applications or commercial objectives in mind.
Of course, it costs money to follow where curiosity leads; scientists need funding to pursue questions about the natural and material world.
About 40% of basic research in the U.S. has been federally funded in recent years. The government makes this investment because basic research is the foundation of long-term innovation, economic growth and societal well-being.
Funding for basic research is distributed by the federal government through several agencies and institutes. For more than a century, the U.S. National Institutes of Health have sponsored a breadth of scientific and health research and education programs. Since 1950, the National Science Foundation has advanced basic research and education programs, including the training of the next generation of scientists.
Other federal agencies have complementary missions, such as the Defense Advanced Research Projects Agency, created in response to the Soviet Union’s launch of Sputnik in 1957. DARPA focuses on technological innovations for national security, many of which have become fixtures of civilian life.
Through a competitive review process at these agencies, subject experts vet research proposals and make funding recommendations. The amount of funding available from the NIH, NSF and DARPA varies annually, depending on congressional appropriations. Most of the awarded funds go to universities, research institutions and other health and science organizations that conduct research. The sum of research dollars awarded differs among states.
Applying research
Scientists undertake basic research to generate new knowledge with no specific end goal in mind. Applied research is different in that it aims to find solutions to real-world problems.
Research that investigates specific, practical objectives or improvements with commercial potential is more likely to attract private investors. Companies directly invest in research and development to gain a competitive edge and turn a profit. Private industry is more likely to sink dollars into applied rather than basic research because the potential payoff in the form of a new product or advance is more visible.
From discovery to real-world implementation
As applied research addresses problems, promising findings are moved toward clinical application or mainstream use. This research and development process can lead to tangible benefits for individuals and society.
According to numbers reported by a coalition of research institutions, every dollar that NIH spends on research leads to $2.56 of new economic activity. For the 2024 fiscal year, this means, of the $47.35 billion Congress appropriated for NIH, the $36.94 billion awarded to U.S. researchers fueled $94 billion in activity through employment and the purchase of research-related goods and services.
Economist Pierre Azoulay and colleagues recently imagined an alternative history where NIH was 40% smaller and dispersed less money – a budget akin to current federal proposals. They argued that more than half of the drugs FDA approved since 2000 are tied to NIH-funded research that would have been cut under this scenario. This thought experiment underscores how valuable those basic research dollars are.
‘Last Week Tonight with John Oliver’ points out some seemingly outlandish basic research that has yielded surprising real-world applications.
Even seemingly out-of-touch or abstract studies may precede discoveries with major impact. Basic research into bee nectar foraging and movement around the colony, recently mentioned on “Last Week Tonight with John Oliver,” led to the development of an algorithm that distributes internet traffic between computer servers, which now powers the multibillion-dollar web-hosting industry. Learning about applications of research with visible societal impacts can help people understand and appreciate the role of funding in the scientific enterprise.
Ryan Summers receives funding from the National Science Foundation (NSF) and the National Institutes of Health (NIH). He is affiliated with the Association for Science Teacher Education (ASTE), NARST, which is a global organization for improving science education through research, and the National Science Teaching Association (NSTA).
How do kids figure out how to sort things by order?Celeste Kidd
I’m in a coffee shop when a young child dumps out his mother’s bag in search of fruit snacks. The contents spill onto the table, bench and floor. It’s a chaotic – but functional – solution to the problem.
Children have a penchant for unconventional thinking that, at first glance, can look disordered. This kind of apparently chaotic behavior served as the inspiration for developmental psychologist Jean Piaget’s best-known theory: that children construct their knowledge through experience and must pass through four sequential stages, the first two of which lack the ability to use structured logic.
Piaget remains the GOAT of developmental psychology. He fundamentally and forever changed the world’s view of children by showing that kids do not enter the world with the same conceptual building blocks as adults, but must construct them through experience. No one before or since has amassed such a catalog of quirky child behaviors that researchers even today can replicate within individual children.
While Piaget was certainly correct in observing that children engage in a host of unusual behaviors, mylab recently uncovered evidence that upends some long-standing assumptions about the limits of children’s logical capabilities that originated with his work. Our new paper in the journal Nature Human Behaviour describes how young children are capable of finding systematic solutions to complex problems without any instruction.
Jean Piaget describes how children of different ages tackle a sorting task, with varying success.
Putting things in order
Throughout the 1960s, Piaget observed that young children rely on clunky trial-and-error methods rather than systematic strategies when attempting to order objects according to some continuous quantitative dimension, like length. For instance, a 4-year-old child asked to organize sticks from shortest to longest will move them around randomly and usually not achieve the desired final order.
Psychologists have interpreted young children’s inefficient behavior in this kind of ordering task – what we call a seriation task – as an indicator that kids can’t use systematic strategies in problem-solving until at least age 7.
Somewhat counterintuitively, my colleagues and I found that increasing the difficulty and cognitive demands of the seriation task actually prompted young children to discover and use algorithmic solutions to solve it.
Piaget’s classic study asked children to put some visible items like wooden sticks in order by height. Huiwen Alex Yang, a psychology Ph.D. candidate who works on computational models of learning in my lab, cranked up the difficulty for our version of the task. With advice from our collaborator Bill Thompson, Yang designed a computer game that required children to use feedback clues to infer the height order of items hidden behind a wall, .
The game asked children to order bunnylike creatures from shortest to tallest by clicking on their sneakers to swap their places. The creatures only changed places if they were in the wrong order; otherwise they stayed put. Because they could only see the bunnies’ shoes and not their heights, children had to rely on logical inference rather than direct observation to solve the task. Yang tested 123 children between the ages of 4 and 10.
Researcher Huiwen Alex Yang tests 8-year-old Miro on the bunny sorting task. The bunnies are hidden behind a wall with only their sneakers visible. Miro’s selections exemplify use of selection sort, a classic efficient sorting algorithm from computer science. Kidd Lab at UC Berkeley.
More than half the children we tested demonstrated evidence of structured algorithmic thinking, and at ages as young as 4 years old. While older kids were more likely to use algorithmic strategies, our finding contrasts with Piaget’s belief that children were incapable of this kind of systematic strategizing before 7 years of age. He thought kids needed to reach what he called the concrete operational stage of development first.
Our results suggest that children are actually capable of spontaneous logical strategy discovery much earlier when circumstances require it. In our task, a trial-and-error strategy could not work because the objects to be ordered were not directly observable; children could not rely on perceptual feedback.
Explaining our results requires a more nuanced interpretation of Piaget’s original data. While children may still favor apparently less logical solutions to problems during the first two Piagetian stages, it’s not because they are incapable of doing otherwise if the situation requires it.
A systematic approach to life
Algorithmic thinking is crucial not only in high-level math classes, but also in everyday life. Imagine that you need to bake two dozen cookies, but your go-to recipe yields only one. You could go through all the steps of making the recipe twice, washing the bowl in between, but you’d never do that because you know that would be inefficient. Instead, you’d double the ingredients and perform each step only once. Algorithmic thinking allows you to identify a systematic way of approaching the need for twice as many cookies that improves the efficiency of your baking.
Algorithmic thinking is an important capacity that’s useful to children as they learn to move and operate in the world – and we now know they have access to these abilities far earlier than psychologists had believed.
That children can engage with algorithmic thinking before formal instruction has important implications for STEM – science, technology, engineering and math –education. Caregivers and educators now need to reconsider when and how they give children the opportunity to tackle more abstract problems and concepts. Knowing that children’s minds are ready for structured problems as early as preschool means we can nurture these abilities earlier in support of stronger math and computational skills.
And have some patience next time you encounter children interacting with the world in ways that are perhaps not super convenient. As you pick up your belongings from a café floor, remember that it’s all part of how children construct their knowledge. Those seemingly chaotic kids are on their way to more obviously logical behavior soon.
Celeste Kidd receives funding from the National Science Foundation, the John Templeton Foundation, the Jacobs Foundation, and the Advanced Research and Invention Agency.
Scientists imagine virtual particles popping in and out of existence to explain how forces transfer between particles. koto_feja/iStock via Getty Images
A clever mathematical tool known as virtual particles unlocks the strange and mysterious inner workings of subatomic particles. What happens to these particles within atoms would stay unexplained without this tool. The calculations using virtual particles predict the bizarre behavior of subatomic particles with such uncanny accuracy that some scientists think “they must really exist.”
Virtual particles are not real – it says so right in their name – but if you want to understand how real particles interact with each other, they are unavoidable. They are essential tools to describe three of the forces found in nature: electromagnetism, and the strong and weak nuclear forces.
Real particles are lumps of energy that can be “seen” or detected by appropriate instruments; this feature is what makes them observable, or real. Virtual particles, on the other hand, are a sophisticated mathematical tool and cannot be seen. Physicist Richard Feynman invented them to describe the interactions between real particles.
But many physicists are not convinced by this cut-and-dried distinction.
Although researchers can’t detect these virtual particles, as tools of calculation they predict many subtle effects that ultrasensitive experiments have confirmed to a mind-boggling 12 decimal places. That precision is like measuring the distance between the North and South poles to better than the width of a single hair.
This level of agreement between measurements and calculations makes virtual particles the most thoroughly vetted idea in science. It forces some physicists to ask: Can a mathematical tool become real?
Virtual particles help scientists follow the interactions between particles.
A bookkeeping tool
Virtual particles are the tool that physicists use to calculate how forces work in the microscopic subatomic world. The forces are real because they can be measured.
But instead of trying to calculate the forces directly, physicists use a bookkeeping system where short-lived virtual particles carry the force. Not only do virtual particles make the calculations more manageable, they also resolve a long-standing problem in physics: How does a force act across empty space?
One big advantage of this tool is that the mathematical operations describing the forces between particles can be visualized as diagrams. They tend to look like stick-figure cartoons of particle pingpong played with virtual particles. The diagrams – dubbed Feynman diagrams – offer an excellent intuitive framework, but they also give virtual particles an aura of reality that is deceiving.
Feynman diagrams help physicists calculate particle interactions.
Amazingly, this virtual particle-based method for calculation produces some of the most precise predictions in all of science.
As a professor of physics and astronomy at Mississippi State University, I perform experiments that often rely on the idea that the electrons and protons seen in our instruments interact by swapping virtual particles. My colleagues and I have recently measured the size of the proton very precisely, by bombarding hydrogen atoms with a beam of electrons. This measurement assumes that the electrons can “feel” the proton at the center of the hydrogen atom by exchanging virtual photons: particles of electromagnetic energy.
Physicists use virtual particles to calculate how two electrons repel each other, with exquisite precision. The forces involved are represented as the accumulated effect of the two electrons trading virtual photons.
When two metal plates are placed extremely close together in a vacuum, they attract each other: This is known as the Casimir effect. Physicists can accurately calculate the force that pulls the plates together using virtual particle mathematics. Whether the virtual particles are really there or not, the math predicts exactly what researchers observe in the real world.
Virtual particles can help explain how black holes act. SXS, CC BY-ND
Yet another mysterious prediction made using the virtual particle tool kit is so-called Hawking radiation. When virtual particle pairs pop into existence at the edge of black holes, sometimes the black hole’s gravity grabs one partner while the other escapes. This rift causes the black hole to slowly evaporate. Although Hawking radiation has not yet been directly observed, researchers have recently observed it indirectly.
Useful fiction
Let’s circle back to the question: Can a mathematical tool become real? If you can perfectly predict everything about a force by imagining it is carried by virtual particles, do these particles qualify as real? Does their fictional status matter?
Physicists remain divided on these questions. Some prefer to “just shut up and calculate” – one of Feynman’s famous quips. For now, virtual particles are our best way to describe how particles behave. But researchers are developing alternative methods that do not need them at all.
If successful, these approaches could make virtual particles vanish for good. Successful or not, the fact that alternatives exist at all suggests virtual particles might be useful fiction rather than physical truth. It also fits the pattern of previous revolutions in science – the example of ether comes to mind. Physicists invented ether as a medium through which light waves traveled. Experiments matched well with calculations using this tool, yet they could not actually detect it. Eventually, Einstein’s theory of relativity showed it was unnecessary.
Virtual particles are a striking paradox of modern physics. They shouldn’t exist, yet they are indispensable for calculating everything from the strength of magnets to the behavior of black holes. They represent a profound dilemma: Sometimes the best insights into reality come through carefully constructed illusion. In the end, confusion around virtual particles may be just the price of understanding fundamental forces.
Dipangkar Dutta receives funding from US Dept. of Energy and NSF.
Tilly Norwood is the hottest actor in Hollywood right now.
Her career has been covered by Variety, the BBC and Forbes, to name just a few publications. All of this is publicity that a young actor at the start of their career can only dream of. But Tilly doesn’t dream. Nor is she actually acting in the strictest sense of the word, because Tilly is an AI actor, created by Particle6 Studios, a UK-based AI-focused film production company.
There have, of course, been AI actors before. Carrie Fisher was famously resurrected for The Rise of Skywalker in 2019. James Cameron used background “actors” to populate Titanic in 1997, but until now no AI creation has achieved the media cut-through that Tilly has. This is partly due to her creator – Eline Van Der Velden – and her team. They have launched Tilly into the marketplace as a persona: something designed to act and emote.
As Van Der Velden told entertainment news site Screen Daily: “[Tilly is] an act of imagination and craftsmanship, not unlike drawing a character, writing a role or shaping a performance.” There is technological craft in her creation, certainly. But there is also a grey area, where that creation draws on the work, voices, physiognomy and artistry of others – blended into code, shaped for modern media and packaged in a soft-focus comedy video just meta enough to deflect criticism.
Tilly Norwood appears in an AI sitcom sketch.
My work with actors has always been deeply rewarding. At the Guilford School of Acting, where I teach, the approach is grounded in the belief that acting is born from a combination of craft, empathy, collaboration and above all a genuine exploration of what it means to be human. The story of “Tilly’s” creation has stirred a powerful response among the students I have been working with: a mix of horror, fear and, perhaps most chillingly of all, resignation. Resignation that this may indeed be the direction in which the creative industries are heading.
The outcry from established actors was immediate and heartfelt. On hearing that agents were already contacting the production company in hopes of representing it, A-lister Emily Blunt told interviewers: “Good lord, we’re screwed. That is really, really scary. Come on, agencies, don’t do that. Please stop. Please stop taking away our human connection.”
The human connection is the point. The Russian theatre practitioner, Konstantin Stanislavski, whose work consistently urged actors to seek inner truth and humanity, summed it up well. Writing in his book An Actor Prepares (1936), he explained: “To break that rule of using your own feelings is the equivalent of killing the person you are portraying, because you deprive him of a palpitating, living, human soul, which is the real source of life for a part.”
In a recent podcast interview with Jay Shetty, actor Emma Watson reflected on how the “movie star” version of herself had become something of an avatar in her mind. She spoke candidly about her journey from the Harry Potter films, the hypersexualisation she endured in the media and the scrutiny now placed on her every word and stance.
For producers, directors, and studios, a compliant, commodified figure like Norwood is an attractive prospect: an actor who doesn’t need an intimacy coordinator, won’t go off-message on social media or perhaps more disturbingly, might. As impressive as the technological achievement is, the choice of an elfin-thin, 20-something female “actor” is also highly questionable.
In a world where power dynamics and abuses are finally being called out through the #MeToo movement, it’s perhaps no surprise that the coded, painted and constructed Tilly Norwood has arrived. The “actor” is programmable and usable. It looks human but is, at its core, deficient. And will always remain so. Because what makes an actor is that ineffable thing: humanity.
Looking for something good? Cut through the noise with a carefully curated selection of the latest releases, live events and exhibitions, straight to your inbox every fortnight, on Fridays. Sign up here.
Nicholas Scrivens does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
Does all music sound the same these days? Many listeners – and artists – think so. There’s a concern that today’s hits are increasingly generic, predictable and indistinguishable. And it might all come down to money.
Streaming platforms like Spotify have transformed music production, distribution and consumption. In place of nurturing individual expression, there’s long been a belief that streaming platforms have shifted the focus to financial goals.
Our new research examined these perceptions and found that over a 20-year period there was a move towards standardisation, repetition and conformity in popular tracks.
In the 1940s, philosophers like Theodor Adorno and Max Horkheimer argued that, much like Henry Ford’s production lines, music had become a mass-produced commodity designed for passive consumption. By the early 2000s, physical record sales still drove revenues, major labels controlled most of the market, and promotional power was concentrated in radio stations, music television and charts.
Despite this commodified structure, however, music – especially in genres like hip-hop – remained stylistically diverse and regionally distinct.
Yet, over the last decade, a transformation has occurred. The rise of streaming platforms and the growing role of finance have restructured the culture industry. This is not only changing how music is distributed, but fundamentally altering how it is valued and produced.
In our study, we found that today’s industry is no longer primarily about selling commodities like albums, tickets or CDs. Rather, it is about generating financial assets in the form of rising numbers of plays and subscriptions that promise to create future income streams.
This shift is driven by two major forces that we call “platformisation” and “finacialisation”. Platformisation refers to the dominance of streaming services that shape how music is produced and consumed. Financialisation is about prioritising future income streams over immediate profitability.
In this new landscape, value is created not by sales but by ownership over future income. This is turning songs, playlists and platforms into financial assets. It has transformed music into an investment product and playlists into highly curated tools for extracting value.
Spotify, for instance, rarely turns a profit. Instead, its business model revolves around an expectation about future increases in revenue. This lies in increasing plays from both paid and unpaid subscriptions, either by increasing advertising revenue or monthly subscription fees.
Investors value Spotify not for its current earnings but for its capacity to grow. To do this, it must maximise plays and subscriptions and “minimise friction” (that is, making the listening experience smooth and uninterrupted). This is where the playlist comes in.
Radio once played a central role in shaping musical tastes. But today, playlists have taken over. With nearly 16 million followers, the highly influential hip-hop playlist RapCaviar does not just reflect listener tastes – it shapes them.
Getting a song on important playlists can generate hundreds of thousands of dollars in revenue, and failing to be listed can mean obscurity. This pressure has changed how music is made.
To be playlisted, songs must conform to a set of unwritten rules: short durations, instant hooks, predictable beats and familiar sonic textures. Songs that deviate too much from the standard risk being skipped and therefore not generating royalties. The result is playlists that are optimised for bingeability and selected for seamless consumption.
To test whether these pressures are leading to the homogenisation of music, we conducted a comparative content analysis of hip-hop music from two eras.
For the pre-streaming period, we examined Apple Music’s retrospective chart playlist of the biggest hip-hop and R&B hits from 2002. For the streaming era, we analysed Spotify’s RapCaviar playlist from 2022.
Both contained a sample of 50 songs that we analysed across five categories. We investigated form and structure, sampling, rhythm, vocal style and lyrics – and the findings were striking.
Song length: the average track duration fell from four minutes and 19 seconds (2002) to three minutes and three seconds (2022), reflecting the pressure to engage listeners quickly
Tempo and key: songs in 2022 clustered much more around similar tempos and harmonic keys, reducing the variety of sound
Samples: where early-2000s tracks drew inspiration from diverse genres and local cultures, most 2022 hits favoured similar moods – generic piano and guitar loops – often sourced from pre-packaged production platforms like LANDR
Rhythm: while earlier hip-hip songs often used distinct rhythms, 90% of 2022 songs used nearly identical 808s (a synthetic drum machine) and rhythms
Vocals: auto-tune effects were nearly ubiquitous in 2022, giving voices a uniform, digital texture
Lyrics: using natural language processing (an AI tool), we found that lyrics in 2022 were 60% more similar to each other than in 2002 – even though they used a larger collection of words.
Taken together, these trends suggest that the sonic and stylistic diversity once praised in hip-hop has been replaced by algorithmic compatibility. While in 2002 a diverse group of songs including Busta Rhymes’ Make It Clap, Eminem’s Lose Yourself or Missy Elliott’s Work It were at the top of hip-hop charts, today’s songs on RapCaviar are much more homogeneous.
Once an art form defined by regionality, resistance and individual expression, hip-hop is increasingly shaped by the incentives of platform capitalism.
Why this matters
This speaks to a broader transformation in how cultural products are made, valued and circulated. Music and other art forms are increasingly produced within platforms designed for scalability. As such, often the asset logic replaces artistic freedom, and predictability trumps originality.
Streaming platforms might claim to democratise the music industry, but in reality they often reinforce the dominance of major labels and pre-existing trends.
Even artists who have benefited from these systems are beginning to speak out about their constraints. This is even more important with the rise of generative AI and the possibility of a future of individualised, on-demand music generation.
If music is to reclaim its critical, creative and expressive power, it needs to be disentangled from the financial logic that now governs it. The first step is understanding how this logic works – and whose interests it serves.
The authors do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.
