As President Donald Trump moved Tuesday to give artificial intelligence a new name — “super intelligence” — Virginia Sen. Mark Warner warned that changing the terminology does nothing to address what he says is the far more consequential question confronting Washington: whether increasingly powerful artificial intelligence systems are advancing faster than the safeguards designed to control them.
“The companies building the most powerful AI systems are warning us that the technology is advancing faster than our safeguards,” Warner, the Democratic vice chairman of the Senate Select Committee on Intelligence, said Tuesday. “The president’s response? To rename it and tell the companies developing it to regulate themselves.”
Trump announced Tuesday that he planned to formally replace “artificial intelligence” with “super intelligence” in federal government documents through an executive order, after previewing the terminology at the United Nations last week. Speaking at the launch of America.gov, the administration’s new AI-powered portal for federal information and services, Trump called “super intelligence,” or “S.I.,” the official name in the United States and said he would sign an executive order implementing the change.
America.gov itself represents how quickly artificial intelligence is moving from experimental technology into everyday government infrastructure. According to the General Services Administration, the platform combines information from more than 29,000 federal government websites and uses AI-enhanced search to provide a single gateway for Americans looking for federal services, benefits and information.
The administration also announced a voluntary agreement with major artificial-intelligence companies involving internal and external reviews of AI risks, part of a broader approach emphasizing American technological development and industry participation rather than the regulatory framework Warner is seeking.
Warner said Tuesday that the distinction is much more than a disagreement over terminology.
“Changing what we call artificial intelligence does nothing to address the very real risks that come with increasingly powerful AI systems,” Warner said.
The Virginia senator emphasized that he remains an advocate for American leadership in artificial intelligence but argued that leadership requires security requirements capable of keeping pace with increasingly powerful models.
“I believe deeply in the promise of AI and in America’s ability to lead the world in this technology,” Warner said. “But leadership also means making sure these systems are safe, secure, and remain under human control.”
Warner’s statement comes just five days after he joined Sens. Brian Schatz (D-Hawaii) and Andy Kim (D-N.J.) in introducing the Artificial Intelligence Risk Management and Security Act of 2026, legislation that would establish enforceable security and testing requirements for the nation’s most advanced artificial-intelligence systems.
The legislation would establish a permanent Artificial Intelligence Safety Board within the Department of Commerce, bringing together representatives from the National Institute of Standards and Technology, Commerce Department, Cybersecurity and Infrastructure Security Agency, National Security Agency and Treasury Department, along with independent technical experts.
Developers of qualifying frontier AI models would be required to provide the board access to those systems at least 45 days before public release, including model weights, configuration files, runtimes and software libraries needed to operate them.
The proposal reaches well beyond the consumer chatbots most Americans now associate with artificial intelligence.
It specifically addresses increasingly capable frontier models, including systems capable of discovering and exploiting software vulnerabilities without direct human prompting. Warner’s legislation would require the AI Safety Board to establish standards for evaluating those capabilities and securing the environments in which the models are tested.
Developers would be legally required to comply with applicable standards, with violations subject to civil penalties of as much as $250,000 per violation, per day.
The bill would also establish a national AI incident database, require companies to report serious safety and security incidents, establish secure federal testing environments and develop standards specifically governing increasingly autonomous AI agents.
Those standards would include something particularly important as artificial intelligence moves beyond answering questions and begins interacting with other computer systems: what information an autonomous AI agent can access, what tools it can use and how much authority it possesses.
Warner’s proposal calls for standards addressing identity, authentication, authorization, access to data and systems and differing levels of autonomy. Developers would also have to document an AI agent’s intended uses, authority boundaries, access to information and tools, known limitations and independent evaluation results where applicable.
The legislation specifically targets advanced systems exhibiting, or capable of being modified to exhibit, high levels of performance on tasks that could pose serious risks to national security, economic security or public health and safety.
Among the risks identified are AI systems facilitating cyberattacks, evading the control of their developers or operators, assisting in the development of chemical, biological, radiological or nuclear weapons, or having their underlying model weights stolen or maliciously modified.
The warnings are not arriving in isolation.
On Sept. 21, Warner joined Sens. Jack Reed (D-R.I.) and Chris Coons (D-Del.) in demanding that the Pentagon and Office of the Director of National Intelligence investigate reported failures involving artificial intelligence in U.S. military and intelligence operations.
Their letter cited public reports involving an AI-targeting platform that allegedly relied on outdated geospatial information during a kinetic strike in Minab, Iran, and a separate military interdiction operation that was aborted after AI-generated false information was included in disseminated U.S. intelligence. The senators asked that inspectors general receive unrestricted access to investigate the reported incidents.
That means the debate underway in Washington is no longer simply about whether artificial intelligence can write an email, generate a photograph or answer a question.
It is increasingly about what happens when artificial intelligence is connected to real systems, real information and real decisions.
That distinction is becoming more important as AI systems move into government operations, cybersecurity, national security and autonomous applications. America.gov itself demonstrates one side of that transition: artificial intelligence is now being placed between Americans and information drawn from tens of thousands of government websites.
Warner is arguing that the most powerful systems require enforceable safeguards before those capabilities increase further.
“There is bipartisan interest in Congress in getting this right,” Warner said Tuesday. “The technology is moving incredibly fast, and the risks are no longer theoretical.”
“The question is whether Washington is going to act before something goes seriously wrong — or wait until afterward to ask why we didn’t.”
That question is becoming increasingly difficult to separate from another one: What happens when increasingly capable artificial intelligence is connected to the enormous amount of information governments and private companies have already collected?
A database by itself is information. A camera records an image. A license-plate reader records a vehicle. Location information can reveal movement. A commercial data broker can assemble information from multiple sources. Social networks contain enormous maps of relationships, interests and behavior. Government databases contain still another universe of information.
For decades, the technical challenge was collecting, storing, searching and connecting those pieces.
Artificial intelligence is rapidly changing the scale at which that can be done.
Washington is now confronting two related questions at once: how powerful these systems should be permitted to become without enforceable safeguards, and what authority they should have when connected to data and other systems.
Trump’s approach emphasizes rapid American technological development and industry-led safeguards. Warner’s legislation would impose mandatory requirements on the most advanced systems, including pre-release access for government evaluators and enforceable standards governing security, testing and autonomous agents.
And while Washington debates whether to call the technology artificial intelligence or super intelligence, the technology itself continues to advance.
EDITOR’S NOTE
Warner’s warning arrives at an extraordinary moment for the Falls Church News-Press.
This week, the News-Press will begin publishing an extensive investigation into the other side of the artificial-intelligence revolution: the massive data and surveillance infrastructure that increasingly powerful AI systems may be capable of searching, connecting and analyzing.
Our investigation did not begin with Warner’s statement Tuesday, and it did not begin with President Trump’s announcement.
It began with a much simpler question:
Who is collecting our data, and where does it go?
Following that question took us backward through more than two decades of government research, intelligence programs, defense contracting, commercial databases and the explosive growth of social media. It took us through the post-9/11 effort to connect previously isolated pieces of information; through government research into enormous datasets and social networks; through the growth of commercial data brokers capable of assembling extraordinarily detailed profiles; and into today’s world of location information, license-plate databases, facial recognition, social-media intelligence and artificial intelligence.
Then the investigation came home.
Again and again, the trail crossed Northern Virginia and the Washington metropolitan region — a region containing the Pentagon, CIA, intelligence and defense contractors, technology companies, data centers and an enormous ecosystem of private businesses serving the federal government.
That matters because the central question surrounding artificial intelligence may no longer be simply how intelligent these systems become.
It may be what we connect them to.
For years, many of the pieces of the modern information economy could be viewed separately. Social media was one debate. Government surveillance was another. Commercial data brokers were another. License-plate readers, location tracking, facial recognition, government databases and artificial intelligence each generated their own arguments over privacy, security, innovation and government power.
But what happens when those pieces are no longer separate?
Artificial intelligence creates the possibility of searching, comparing and analyzing information at a scale and speed no individual human analyst could realistically reproduce. Increasingly autonomous systems raise another question: whether machines will simply find and analyze that information or eventually be permitted to act upon it.
That is precisely where Warner’s warning Tuesday intersects with what the News-Press has been investigating.
His proposed legislation does not merely address chatbots generating text. It specifically addresses frontier models, cybersecurity, autonomous AI agents, access to data and other computer systems, authority boundaries and maintaining human control as those capabilities increase.
At the same time, the federal government Tuesday unveiled an AI-powered platform drawing upon information across more than 29,000 government websites.
Those developments do not establish that the systems examined in our forthcoming investigation are connected to America.gov or to one another. But they demonstrate why the larger question can no longer be dismissed as science fiction.
The technology capable of analyzing enormous quantities of information is becoming more powerful.
The information infrastructure already exists.
And increasingly, the debate is becoming about what happens when the two meet.
Beginning this week, the Falls Church News-Press will examine that infrastructure: where portions of it came from, the government programs and private companies that helped build it, the commercial databases and surveillance technologies operating today, how federal agencies obtain information from private sources, and why so many roads in this story lead directly back to Northern Virginia.
For years, the public has been shown individual pieces of the puzzle.
Facebook. Data brokers. Government surveillance. Location tracking. License-plate readers. Facial recognition. Artificial intelligence. Intelligence agencies. Government contractors.
Our investigation asks what the picture looks like when those pieces are placed together.
Warner says the risks surrounding increasingly powerful artificial intelligence are “no longer theoretical.”
This week, we’re going to show our readers what has already been built — and ask what happens when we connect it all.
