Opinion | The costs of powering the AI revolution

The administration’s plan fails to protect working people from the potential harms of AI, some of which we’re already seeing today. It also invites corporations to flood U.S. markets with experimental AI products that threaten jobs, safety, civil rights and our privacy.

Notably, a wrongheaded proposal to deter states from regulating artificial intelligence for a decade was recently resoundingly rejected by 99 senators.

America’s unions are pushing for a path forward on AI that incorporates workers’ expertise and respects their rights so the benefits of the technology can be enjoyed by all. If President Donald Trump truly wants to empower workers, as his plan announced, he should work with unions to achieve that goal.

The writer is interim director of the AFL-CIO Technology Institute.

President Donald Trump’s new “AI Action Plan” looks to expand American innovation in the field of artificial intelligence.

It’s a laudable goal. However, to achieve these aims, the Trump administration must invest proper resources in the entire ecosystem that drives the basic scientific research needed to power new AI breakthroughs. That includes investing in colleges and universities across the country. We should not wait for private research investment that might never come. The Trump administration should treat AI research as a priority, but its recent cuts to federally funded university research put our nation’s status as a leader in AI development at risk.

Academic research, backed by robust federal investments and put into action by the private sector, is crucial to winning the AI race. If the United States’ goal is to remain at the forefront of the world’s AI race, lawmakers must reverse recent funding cuts to research at U.S. colleges and universities. Without these vital investments, we will cede our AI dominance to foreign adversaries such as China, and we will lose this critical competition.

Abigail Robbins, Washington

The writer is president of the Science Coalition.

As was correctly stated in the July 28 front-page article “Boom in AI drives higher electric bills for millions,” data centers are consuming more energy as these facilities support the ever-growing artificial intelligence and cloud computing industries. The increases in electrical consumption impact us all through utility costs.

This article reminded me of one issue that I have always wondered about: What would be the impacts of all Americans switching to electric vehicles? That change would surely create a significant electrical demand. I am fully supportive of technologies that reduce greenhouse gas emissions, but electric cars still require energy production to charge the batteries. If energy prices are going up to support the computing needs of AI, can you imagine the amount of extra energy production needed to power electric cars for the entire country? If we don’t do something to counter our energy needs, I fear we will see even more drastic utility cost increases in the near future.

Adrian Fremont, Alexandria

Unfortunately, reading the July 28 editorial, the July 28 front-page article on artificial intelligence and the July 28 news article “Why extreme weather is so tough on mobile homes” in succession painted a frightening picture of our climate predicament and our apparently deep resolve to make it much worse. We are seeing a dramatic rise in flooding from extreme rain events that claim lives and property, and the explosion of AI data centers is going to ramp up our electricity bills, especially here in Virginia.

I felt dubious of the editorial’s idea that President Donald Trump’s plan for AI is “promising.” Americans face rising waters, rising electricity rates and rising insurance rates. Will AI provide advice that will fix the cumulative climate mess it helps create? I don’t think so. AI might be able to think for us, but it comes with great cost.

To mitigate the effects of this changing climate, we need to cut carbon emissions. Our grandchildren will marvel at the hypocrisies of this time.

Chris Wiegard, Chester, Virginia

Mitch Daniels’s speculations in his July 30 column, “What the heck would you put in a time capsule in 2025?,” about what the world could look like 100 years from now are insightful and thought-provoking. However, he was mistaken to suppose that readers a century from now will shake their heads over how today’s leaders have dealt with climate change, “spending trillions without moving, or any real prospect of moving, the world’s thermometer.” In fact, today’s policies have already moved the thermometer by quite a lot.

The 2014 U.N. Intergovernmental Panel on Climate Change climate report, which was released before the 2015 Paris accord, said the Earth was on track for a global temperature rise of 3.7 to 4.8 degrees Celsius (or 6.7 to 8.6 degrees Fahrenheit) by the year 2100. The latest report, released in 2023, put the likely range between 2.1 and 3.4 degrees Celsius (or 3.8 to 6.1 degrees Fahrenheit). In less than 10 years, we’ve lowered the temperature of our future world by about 1.5 degrees Celsius.

To be clear, this lower level of global temperature rise is still very bad. But it’s not as bad as it could be. In other words, what we’ve been doing over the past decade, such as moving away from fossil fuels and investing in renewable energy, is starting to work. To protect our future, we need to keep up this momentum, not change course.

Amy Livingston, Highland Park, New Jersey

Josh Max’s July 28 online op-ed, “845,000 dead on U.S. highways. Why not address the main cause?,” contended that more driver training would lead to fewer motor vehicle fatalities. Significant obstacles prevent this from being a feasible solution. First, driver education courses are usually only available for those who can afford to pay for them. Second, mass implementation of these programs as requirements would have to be funded from government coffers. Moreover, research doesn’t indicate driver education courses for high-schoolers are effective.

Newer vehicles are often equipped with advanced crash-avoidance features. But these technologies need to be required and subject to minimum safety standards. Otherwise, they will continue to be offered at a cost and will vary in effectiveness.

Last year, the U.S. Transportation Department’s National Highway Traffic Safety Administration prudently issued a requirement that automatic emergency braking with pedestrian detection be installed on new passenger vehicles in the coming years. AEB can mitigate or reduce crashes involving myriad human behavior-related issues. Transportation officials estimate that this action will prevent at least 360 deaths and 24,000 injuries annually. They should expand the requirement to detect bicycle and motorcycle riders. Similarly, installing impaired driving prevention technology — which, according to the Insurance Institute for Highway Safety, could save more than 10,000 lives annually once widely deployed — should be required.

In addition to federal action, states should use speed and red-light safety cameras, require all passengers to buckle up, and require motorcycle riders to wear a helmet. All-offender ignition interlock device laws, which reduce repeat drunken driving offenses, should be the norm in all states. So should more comprehensive graduated driver licensing programs, designed for novice drivers to get the experience that research shows offers one of the best options to reduce their risk of a crash.

Major reductions to our horrific motor-vehicle death and injury toll can be achieved, but our leaders must advance proven solutions expeditiously.

The writer is president of Advocates for Highway and Auto Safety.

Josh Max’s July 28 online op-ed discussed possible reasons behind car accidents, but it should have also mentioned how many new cars are built with huge touch screens in front of the driver. They usually require a focus on the screen, which takes away from the driver’s focus on the road. These screens have become widespread, to the detriment of highway safety. We should return to automobile levers and switches, with no need to take eyes off the road.

Michael Harper, Folsom, California

The headline of Sarah Labowitz’s July 26 op-ed, “FEMA has become its own disaster,” should have been “Trump has become a disaster for FEMA.”

The process that Labowitz described was not the one that occurred when I was at FEMA, and I believe she confused the process for a state to receive federal assistance in a disaster. The 1988 Robert T. Stafford Federal Disaster and Emergency Assistance Act established a system that requires the president to make a federal emergency declaration or a federal disaster declaration. FEMA makes a recommendation, prepares the paperwork, sends it to the Department of Homeland Security and then to the president.

Except for minor items after the president has made the declaration, FEMA does not issue partial approvals or disapprovals. The statute and regulations require the president to make the ultimate decision. FEMA merely makes a recommendation to the president. The apparent practice of “batching” requests and recommendations of FEMA professionals, which now go through DHS and then to the president, makes a mockery of the sovereignty of the individual states and their governors and has never been FEMA’s practice.

FEMA’s long-standing tradition treats each state as sovereign and individual.

Diane L. Donley, Alexandria

The writer is a retired Federal Emergency Management Agency attorney.