Category: Analysis

Source: The Conversation – USA – By Carlos Gershenson, Professor of Innovation, Binghamton University, State University of New York

In the past decade, AI’s success has led to uncurbed enthusiasm and bold claims – even though users frequently experience errors that AI makes. An AI-powered digital assistant can misunderstand someone’s speech in embarrassing ways, a chatbot could hallucinate facts, or, as I experienced, an AI-based navigation tool might even guide drivers through a corn field – all without registering the errors.

People tolerate these mistakes because the technology makes certain tasks more efficient. Increasingly, however, proponents are advocating the use of AI – sometimes with limited human supervision – in fields where mistakes have high cost, such as health care. For example, a bill introduced in the U.S. House of Representatives in early 2025 would allow AI systems to prescribe medications autonomously. Health researchers as well as lawmakers since then have debated whether such prescribing would be feasible or advisable.

How exactly such prescribing would work if this or similar legislation passes remains to be seen. But it raises the stakes for how many errors AI developers can allow their tools to make and what the consequences would be if those tools led to negative outcomes – even patient deaths.

As a researcher studying complex systems, I investigate how different components of a system interact to produce unpredictable outcomes. Part of my work focuses on exploring the limits of science – and, more specifically, of AI.

Over the past 25 years I have worked on projects including traffic light coordination, improving bureaucracies and tax evasion detection. Even when these systems can be highly effective, they are never perfect.

For AI in particular, errors might be an inescapable consequence of how the systems work. My lab’s research suggests that particular properties of the data used to train AI models play a role. This is unlikely to change, regardless of how much time, effort and funding researchers direct at improving AI models.

Nobody – and nothing, not even AI – is perfect

As Alan Turing, considered the father of computer science, once said: “If a machine is expected to be infallible, it cannot also be intelligent.” This is because learning is an essential part of intelligence, and people usually learn from mistakes. I see this tug-of-war between intelligence and infallibility at play in my research.

In a study published in July 2025, my colleagues and I showed that perfectly organizing certain datasets into clear categories may be impossible. In other words, there may be a minimum amount of errors that a given dataset produces, simply because of the fact that elements of many categories overlap. For some datasets – the core underpinning of many AI systems – AI will not perform better than chance.

For example, a model trained on a dataset of millions of dogs that logs only their age, weight and height will probably distinguish Chihuahuas from Great Danes with perfect accuracy. But it may make mistakes in telling apart an Alaskan malamute and a Doberman pinscher, since different individuals of different species might fall within the same age, weight and height ranges.

This categorizing is called classifiability, and my students and I started studying it in 2021. Using data from more than half a million students who attended the Universidad Nacional Autónoma de México between 2008 and 2020, we wanted to solve a seemingly simple problem. Could we use an AI algorithm to predict which students would finish their university degrees on time – that is, within three, four or five years of starting their studies, depending on the major?

We tested several popular algorithms that are used for classification in AI and also developed our own. No algorithm was perfect; the best ones − even one we developed specifically for this task − achieved an accuracy rate of about 80%, meaning that at least 1 in 5 students were misclassified. We realized that many students were identical in terms of grades, age, gender, socioeconomic status and other features – yet some would finish on time, and some would not. Under these circumstances, no algorithm would be able to make perfect predictions.

You might think that more data would improve predictability, but this usually comes with diminishing returns. This means that, for example, for each increase in accuracy of 1%, you might need 100 times the data. Thus, we would never have enough students to significantly improve our model’s performance.

Additionally, many unpredictable turns in lives of students and their families – unemployment, death, pregnancy – might occur after their first year at university, likely affecting whether they finish on time. So even with an infinite number of students, our predictions would still give errors.

The limits of prediction

To put it more generally, what limits prediction is complexity. The word complexity comes from the Latin plexus, which means intertwined. The components that make up a complex system are intertwined, and it’s the interactions between them that determine what happens to them and how they behave.

Thus, studying elements of the system in isolation would probably yield misleading insights about them – as well as about the system as a whole.

Take, for example, a car traveling in a city. Knowing the speed at which it drives, it’s theoretically possible to predict where it will end up at a particular time. But in real traffic, its speed will depend on interactions with other vehicles on the road. Since the details of these interactions emerge in the moment and cannot be known in advance, precisely predicting what happens to the the car is possible only a few minutes into the future.

Not with my health

These same principles apply to prescribing medications. Different conditions and diseases can have the same symptoms, and people with the same condition or disease may exhibit different symptoms. For example, fever can be caused by a respiratory illness or a digestive one. And a cold might cause cough, but not always.

This means that health care datasets have significant overlaps that would prevent AI from being error-free.

Certainly, humans also make errors. But when AI misdiagnoses a patient, as it surely will, the situation falls into a legal limbo. It’s not clear who or what would be responsible if a patient were hurt. Pharmaceutical companies? Software developers? Insurance agencies? Pharmacies?

In many contexts, neither humans nor machines are the best option for a given task. “Centaurs,” or “hybrid intelligence” – that is, a combination of humans and machines – tend to be better than each on their own. A doctor could certainly use AI to decide potential drugs to use for different patients, depending on their medical history, physiological details and genetic makeup. Researchers are already exploring this approach in precision medicine.

But common sense and the precautionary principle
suggest that it is too early for AI to prescribe drugs without human oversight. And the fact that mistakes may be baked into the technology could mean that where human health is at stake, human supervision will always be necessary.

Carlos Gershenson 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.

– ref. AI’s errors may be impossible to eliminate – what that means for its use in health care – https://theconversation.com/ais-errors-may-be-impossible-to-eliminate-what-that-means-for-its-use-in-health-care-251036

Source: The Conversation – USA – By Lisa Fazio, Associate Professor of Psychology, Vanderbilt University

Zohran Mamdani as a creepy trick-or-treater, Gavin Newsom body-slamming Donald Trump and Hakeem Jeffries in a sombrero. This is not the setup to an elaborate joke. Instead, these are all examples of recent AI-generated political videos. New easy-to-use tools – and acceptance of those tools by politicians – means that these fake videos are quickly becoming commonplace in American politics.

Perhaps the most interesting thing about many of the videos is how clearly fake they are. Rather than trying to deceive the viewer into thinking a depicted event actually happened, the videos serve a different purpose. President Trump didn’t post a video of himself wearing a crown in a fighter jet dumping feces on a group of protesters because he wanted people to believe that the flight actually happened. He likely did it to express his feelings about the protest and to create an in-joke with his followers.

Fears about the political implications of AI-generated videos have been around since the term deepfakes was coined in 2017. Steady improvements in the technology mean that distinguishing real from fake could become a significant threat. But today’s use of AI imagery is largely about making memes and making money – in other words, typical social media content.

Getting a rise out of people

Internet platforms use algorithms designed to keep people engaged, and that typically means promoting content that stirs emotions. AI-generated political videos often provoke an emotional response – amusement or outrage.

People are more likely to share information when it is emotionally arousing. For example, people are more likely to pass along urban legends that elicit feelings of disgust, and news articles that are emotionally charged are more likely to make the New York Times list of most emailed articles. Similar patterns occur online, where emotional content is much more likely to go viral than nonemotional content.

In addition, strong emotions can interfere with people’s ability to detect false information. People are worse at distinguishing between true and false political news headlines when they are experiencing stronger emotions – for instance, enthusiasm, excitement or fear. Thus, emotionally appealing AI-generated videos are both more likely to spread and reduce people’s ability to judge whether they are real or fake.

Online politics

Creating and sharing AI videos is also a powerful way for people to demonstrate their allegiances and show their political identities. “I am a Trump supporter, so I post AI videos of ICE detainees crying to own the libs” or “I am a Democrat and so I share Governor Newsom’s AI-video of JD Vance talking about couches to show that I’m in on the joke.”

What’s new in recent months is that campaigns and politicians are using AI-created videos, not just their supporters. An analysis from The New York Times showed that Trump commonly uses AI imagery to “attack enemies and rouse supporters”.

These new tools also allow for active participation in the political process. Rather than simply watching politicians and voting, citizens can play an active role in shaping the conversation between elections.

Information and technology researcher Kate Starbird has written about similar dynamics in the ways that everyday Americans found “evidence” for voter fraud in the 2020 election. Politicians told the public that voter fraud was going to occur, and then when voters saw things that they did not understand when voting, such as the use of Sharpie pens to mark ballots, they interpreted that action as evidence of voter fraud. Politicians then circulated that evidence online to support the false narrative.

New AI tools make this cycle of participatory disinformation even simpler. Instead of reinterpreting actual events as evidence for a false claim, people can easily generate that evidence themselves.

AI video at volume

AI video creation tools make it incredibly easy for people to churn out hundreds of videos, post them online and simply see what content becomes popular and goes viral. In fact, that’s exactly what seems to have happened with recent AI-generated videos of raids by Immigration and Customs Enforcement. According to an investigation by 404 media, Facebook user “USA Journey 897” used to post a variety of real videos of police activity as well as absurd AI videos of people carrying whales and riding tigers.

However, after the release of a new version of OpenAI’s Sora video generator on Sept. 30, 2025, the account switched entirely to posting multiple fake videos of deportations every day. Most of the videos accumulated hundreds of thousands of views, and one fake video of a Walmart employee being detained had over 4 million views.

Typically these accounts are hosted overseas and exist to earn money through creator incentive programs. These incentives create an environment where social media no longer informs people about the world, but instead serves as a fun-house mirror, presenting back to us the world that we want to see – or at least the version of the world that will capture our attention and outrage.

Flowing into the internet

It’s not always easy for people to detect which videos are real and which are AI-generated. A recent audit by the publication Indicator found that platforms regularly fail to properly label AI content. Researchers posted over 500 AI-generated images and videos across Instagram, LinkedIn, Pinterest, TikTok and YouTube. Less than one-third were properly labeled as AI-generated, and even posts generated by the platform’s own AI tools were often missed.

For years, the great fear concerning political deepfakes was that they were going to fool people into believing something happened that didn’t. They still might, but at the moment, AI-generated political videos are a mix of entertainment and memes, legitimate attempts at persuasion, and ways of capturing attention for money.

In other words, they are now just like the rest of the internet. Most of what we see and share is meant to entertain, some is meant to inform and persuade, and a great deal exists solely to monetize our attention.

Lisa Fazio 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.

– ref. AI-generated political videos are more about memes and money than persuading and deceiving – https://theconversation.com/ai-generated-political-videos-are-more-about-memes-and-money-than-persuading-and-deceiving-268977

Source: The Conversation – USA (2) – By Todd L. Pittinsky, Professor of Technology and Society, Stony Brook University (The State University of New York)

Most of the media coverage of the federal government’s recent cuts in federal research money for universities has focused on its effects on a handful of elite Ivy League universities, such as Harvard, Columbia and Cornell.

“When you take money away from a Columbia or a Harvard or other institutions, you’ve just taken away funds from the best researchers,” Toby Smith, the senior vice president for government relations at the Association of American Universities, told CNN in April 2025.

But these schools account for only a small fraction of the nation’s scientific output that federal research money helps generate.

In my view, too many policy discussions and debates obsess over what happens on the campuses of elite colleges. Meanwhile, public universities quietly power the nation’s research engine.

The Ivies do play a critical role in advanced research. But the nation’s public universities make up the backbone of U.S. innovation – research powerhouses such as the University of Michigan, the University of Texas at Austin, Georgia Institute of Technology and Stony Brook University, where I teach.

These places train the overwhelming majority of science, technology, engineering and mathematics graduates for the U.S. workforce and run the lion’s share of federally funded science and engineering research.

Slashing research and development

U.S. colleges and universities spend more than US$108.8 billion annually on research and development, of which about 55% – roughly $60 billion – comes from federal funding via agencies such as the National Institutes of Health and the National Science Foundation. Together, the country’s eight Ivy League schools received approximately $4.6 billion of federal university research and development funding in 2023 – or 7.8% of all federal research and development funding allocated to academia.

In 2023, meanwhile, the University of Washington, Georgia Institute of Technology, the University of California, San Diego, and the University of Michigan, Ann Arbor, each received over $1 billion in federal research funds.

The Trump administration’s federal research and development funding cuts are largely tied to what are called “indirect costs.”

Direct costs fund researcher salaries and lab supplies. Indirect costs support the infrastructure that makes research possible and compliant with federal guidelines: lighting, heating and cooling for labs; high-speed data networks; security; and administrative staff who handle payroll and ensure adherence to federal safety and ethics standards.

In 2025, the Trump administration decided to cap indirect costs for grants awarded by the National Institutes of Health and the National Science Foundation at 15% of the total grant. Traditionally, universities have negotiated their own rates based on documented overhead expenses, with many institutions citing indirect cost rates between 50% and 70%.

Many aspects of the 15% cap have been paused as they’re being challenged in court. Should these cuts go through, it would be incredibly disruptive for universities that have counted on this funding.

However, outside of the battle over indirect costs, many research projects simply lost funding or experienced major delays. Over the past year, thousands of grants have been frozen, terminated or left unfunded.

Ivy League institutions are much better equipped to weather the storm.

In 2021, Forbes reported that the collective endowment of the eight Ivies was approximately $192.6 billion – led by Harvard’s $53.2 billion and Yale’s $42.3 billion. Supporters of Trump’s funding cuts have argued that this immense, tax-exempt financial arsenal could significantly subsidize their overhead costs, rather than relying on taxpayers to do so. While endowments don’t serve as a blank check, schools can still pull from them in times of need.

In contrast, public universities are far more dependent on federal funds to sustain labs, staff and graduate programs. In 2021, the entire Texas public university system – the University of Texas at Austin, Texas A&M, the University of Houston, the University of Texas at Dallas and Texas Tech – held an endowment of around $40 billion, more than $10 billion less than Harvard’s.

Public schools are the training ground

This isn’t to argue that the Ivies should have their research funding cut, while public universities ought to be spared. It’s to shift the focus of the conversation to who stands to lose the most: the public universities that educate the vast majority of the U.S.’s future scientists and fuel most of the nation’s scientific output.

The Ivy League’s geographic reach is extremely concentrated, situated across just seven states, all in the Northeast. Public four-year institutions are located in all 50 states and draw from a much more economically and racially diverse population. They award the vast majority of engineering degrees in the U.S., with more than 144,701 given out in 2023, or more than 70% of the nation’s total.

Purdue University awarded 3,827 engineering degrees that year, with Texas A&M conferring 3,704. By contrast, Cornell University granted just 820 engineering degrees – the most among the Ivies, but just 25th nationally.

Elite schools, including the Ivies, have increasingly steered graduates into finance, law or consulting. Just 2.72% of Yale’s 2024 graduating class was employed as engineers six months after graduation. Meanwhile, public universities serve as the top feeder schools for major defense and aerospace firms.

Stony Brook University’s College of Engineering and Applied Sciences enrolled over 5,600 students in fall 2024, making it one of the top producers of engineering talent in New York. The university manages the Department of Energy’s Brookhaven National Laboratory, which uses advanced tools such as particle colliders to make discoveries in physics, energy, materials and biology. It’s one of a handful of universities that directly operate a national lab.

Collateral damage

The Trump administration has argued that federal research funding cuts are necessary because too many dollars have been allocated for social policy goals, whether it’s $349,985 to train engineers to “enable engineering for social justice” or $600,000 to teach aerospace engineering students “critical consciousness and sensitivity to injustices within social systems.”

While federal law bars the Department of Education from influencing curriculum, the National Science Foundation faces no such constraint. Too much funding, it claimed, went to research that strayed from the federal agencies’ core scientific mandates and crowded out the kind of critical research that underpins U.S. innovation.

While I think fiscal responsibility and mission creep merit attention, cuts that are too dramatic and too indiscriminate risk gutting the high-impact research that is essential to national security and technological leadership, much of which takes place at public institutions. Fields from supercomputing and wireless communications to biothreat countermeasures and health sciences will and have felt the pain of widespread cuts.

Federal research funding is not merely academic spending. It is an important national investment that directly fuels local and national economic growth. It supports high-wage jobs and “innovation districts” centered on university and federal laboratories.

The future of scientific research funding isn’t a debate over how much government help the privileged Ivies ought to receive.

It’s a question of whether public universities will be given the resources to nurture the next generation of scientists, drive bold discoveries and keep America at the forefront of scientific innovation for generations to come.

Todd L. Pittinsky is a Professor of Technology, AI & Society in the College of Engineering and Appled Sciences at Stony Brook University (SUNY).

– ref. The Ivies can weather the Trump administration’s research cuts – it’s the nation’s public universities that have the most to lose – https://theconversation.com/the-ivies-can-weather-the-trump-administrations-research-cuts-its-the-nations-public-universities-that-have-the-most-to-lose-267197

Source: The Conversation – USA (2) – By Kelly Droege, Assistant Professor, University of Wisconsin-Stout

For decades, a four-year college degree was widely seen as the standard path to getting most midlevel jobs in the United States. It was the expected entry point for getting a job as a marketing specialist, project manager, IT support analyst, among other roles.

But this expectation is shifting. Many fields – including cybersecurity, health care and advanced manufacturing – are facing significant shortages in skilled workers. The gap between available skilled jobs and workers is likely to push employers to rethink what they require from job candidates over the next decade.

A major demographic shift will also play a role. Between 2024 and 2032, an estimated 18.4 million experienced workers with education beyond high school are expected to retire, according to September 2025 findings by Georgetown University’s center on education and the workforce.

Only 13.8 million younger workers with similar education levels are expected to enter the workforce during the same period, these findings show. This trend will also make it harder for employers to fill roles that traditionally require a college degree.

At the same time, 25 states over the past few years have enacted legislation and executive orders to remove college degree requirements for people applying for some public sector jobs, signaling a shift in how essential college degrees are for getting hired for some kinds of work.

These shifts underscore a broader trend: A four-year degree is no longer essential for many kinds of work.

Hiring data tells a similar story. As of January 2024, 52% of U.S. job postings on Indeed did not mention any formal education requirement, up from 48% in 2019. Job postings requiring at least a bachelor’s degree also dropped from 20.4% to 17.8% between 2018 and 2023.

As hiring expectations change – influenced in part by advances in artificial intelligence – employers may struggle to find candidates who already have the right job-specific skills.

With over 20 years of experience as professors who also train employees in industries such as manufacturing, health care and business information technologies, we believe that college degrees shouldn’t be mandatory for some jobs.

A widening gap

Nearly half of recent college graduates say they feel unprepared for entry-level work, and 56% cite a lack of job-specific skills as the biggest issue, according to a 2025 report by Cengage Group, an education and workforce training company.

Alternative pathways – apprenticeships, certifications and on-the-job training – can give workers practical skills and help employers fill crucial roles more quickly.

Employers dropping degree requirements is only one step toward this goal. We think it is also important that prospective college students and their families are aware of educational opportunities besides a traditional four-year degree.

Understanding polytechnic universities

Some people think of higher education in terms of traditional liberal arts colleges or research universities. But there are also polytechnic universities, which focus on hands-on, career-aligned learning and emphasize strong STEM and technical programs. These schools often prepare students for exactly the kinds of jobs employers struggle to fill.

There are about 10 major polytechnic universities in the U.S. Some well-known polytechnic universities are California Polytechnic State University in San Luis Obispo, California, Worcester Polytechnic Institute in Worcester, Massachusetts, and State University of New York Polytechnic Institute in Marcy and Albany, New York.

Instead of offering a wide range of liberal arts majors, polytechnic universities offer majors such as engineering, robotics, construction management and information technology.

A central feature of these schools is applied learning – hands-on labs, real-world projects and problem-solving experiences.

Polytechnic students can earn bachelor’s and master’s degrees, but they also can often get short-term certificates in fields such as human resources, instructional design, project management and digital marketing. Many programs also include apprenticeships, such as workplace training specialists.

Students can also pursue microcredentials, which involve short course sequences that build targeted skills, such as business writing or engineering mechanics. These options give students more flexible and affordable ways to learn without committing to a traditional four-year degree.

Polytechnic universities also tend to cost less than research universities and private colleges, and students can use federal financial aid or private loans to attend.

There are some limitations. Polytechnic schools generally offer fewer majors, usually within STEM fields. Their alumni networks may be smaller, and we have found that some people perceive them as less prestigious than traditional universities because they focus more on teaching than on research.

Real world relevance

In March 2025, we asked 10 online instructors at different polytechnic universities how they bring career-focused learning into their classes.

Our research, which will likely be published in 2026, shows that every instructor tried to make their courses feel relevant to real workplaces.

Some instructors used simulations in the course. Others shared examples from their own industry backgrounds with students. All agreed that students learn best when they can clearly connect their coursework to their career goals.

One of the most effective strategies is hiring instructors with deep industry experience. Their professional networks help programs stay aligned with the skills employers currently value.

Not every college wants to become a polytechnic, and not every student wants that style of education.

However, traditional universities can still learn from this model by adding more applied learning, embedding essential job skills into their programs, and partnering more closely with industry. These changes can better prepare students to succeed in the workforce.

Kelly Droege works for the University of Wisconsin – Stout, a polytechnic institution.

Laura Reisinger 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.

– ref. Polytechnic universities focus on practical, career-oriented skills, offering an alternative to traditional universities – https://theconversation.com/polytechnic-universities-focus-on-practical-career-oriented-skills-offering-an-alternative-to-traditional-universities-268349

Source: The Conversation – USA (2) – By Hannah Wiseman, Professor of Law, Penn State

As projections of U.S. electricity demand rise sharply, President Donald Trump is looking to coal – historically a dominant force in the U.S. energy economy – as a key part of the solution.

In an April 2025 executive order, for instance, Trump used emergency powers to direct the Department of Energy to order the owners of coal-fired power plants that were slated to be shut down to keep the plants running.

He also directed federal agencies to “identify coal resources on Federal lands” and ease the process for leasing and mining coal on those lands. In addition, he issued orders to exclude coal-related projects from environmental reviews, promote coal exports and potentially subsidize the production of coal as a national security resource.

But there remain limits to the president’s power to slow the declining use of coal in the U.S. And while efforts continue to overcome these limits and prop up coal, mining coal remains an ongoing danger to workers: In 2025, there have been five coal-mining deaths in West Virginia and at least two others elsewhere in the U.S.

A long legacy

Until 2015, coal-fired power plants generated more electricity than any other type of fuel in the U.S. But with the rapid expansion of a new type of hydraulic fracturing, natural gas became a cheap and stable source for power generation. The prices of solar and wind power also dropped steadily. These alternatives ultimately overcame coal in the U.S. power supply.

Before this change, coal mining defined the economy and culture of many U.S. towns – and some states and regions, such as Wyoming and Appalachia – for decades. And in many small towns, coal-related businesses, including power plants, were key employers.

Coal has both benefits and drawbacks. It provides a reliable fuel source for electricity that can be piled up on-site at power plants without needing a tank or underground facility for storage.

But it’s dirty: Thousands of coal miners developed a disease called black lung. The federal government pays for medical care for some sick miners and makes monthly payments to family members of miners who die prematurely. Burning coal also emits multiple air pollutants, prematurely killing half a million people in the United States from 1999 through 2020.

Coal is dangerous for workers, too. Some coal-mining companies have had abysmal safety records, leading to miner deaths, such as the recent drowning of a miner in a sudden flood in a West Virginia mine. Safety reforms have been implemented since the Big Branch Mine explosion in 2010, and coal miner deaths in the U.S. have since declined. But coal mining remains a hazardous job.

A champion of coal

In both of his terms, Trump has championed the revival of coal. In 2017, for example, Trump’s Department of Energy asked the Federal Energy Regulatory Commission to pay coal and nuclear plants higher rates than the competitive market would pay, saying they were key to keeping the U.S. electricity grid running. The commission declined.

In his second term, Trump is more broadly using powers granted to the president in emergencies, and he is seeking to subsidize coal across the board – in mining, power plants and exports.

At least some of the urgency is coming from the rapid construction of data centers for artificial intelligence, which the Trump administration champions. Many individual data centers use as much power as a small or medium city. There’s enough generation capacity to power them, though only by activating power plants that are idle most of the time and that operate only during peak demand periods. Using those plants would require data centers to reduce their electricity use during those peaks – which it’s not clear they would agree to do.

So many data centers, desperate for 24/7 electricity, are relying on old coal-fired power plants – buying electricity from plants that otherwise would be shutting down.

Limits remain

Despite the Trump adminstration’s efforts to rapidly expand data centers and coal to power them, coal is more expensive than most other fuels for power generation, with costs still rising.

Half of U.S. coal mines have closed within the past two decades, and productivity at the remaining mines is declining due to a variety of factors, such as rising mining costs, environmental regulation and competition from cheaper sources. Coal exports have also seen declines in the midst of the tariff wars.

The U.S. Department of the Interior’s recent effort to follow Trump’s orders and lease more coal on federal lands received only one bid – at a historically low price of less than a penny per ton. But in fact, even if the government gave its coal away for free, it would still make more economic sense for utilities to build power plants that use other fuels. This is due to the high cost of running old coal plants as compared to new natural gas and renewable infrastructure.

Natural gas is cheaper – and, in some places, so are renewable energy and battery storage. Government efforts to prevent the retirement of coal-fired power plants and boost the demand for coal may slow coal’s decline in the short term. In the long term, however, coal faces a very uncertain future as a part of the U.S. electricity mix.

Hannah Wiseman is affiliated with the Center for Progressive Reform. Along with a team of other Penn State researchers, she also received a seed grant from the Penn State Institute of Energy and the Environment for a project entitled “Assessing distributional effects of coal-fired power plant operations on pollution and health.”

Seth Blumsack receives funding from the U.S. National Science Foundation, NASA, the Alfred P. Sloan Foundation and the Heising Simons Foundation.

– ref. Even with Trump’s support, coal power remains expensive – and dangerous – https://theconversation.com/even-with-trumps-support-coal-power-remains-expensive-and-dangerous-269668