Tech bills of the week: Creating an AI-focused agency; Reviewing AI-assisted cyber attacks; and more

Banning artificial superintelligence 

House and Senate lawmakers introduced joint legislation on Wednesday that would place restrictions on the development of more advanced AI tools until a new regulatory agency is created to oversee the creation of such capabilities. 

The proposal — introduced in the House by Rep. Greg Casar, D-Texas, and in the upper chamber by Sen. Bernie Sanders, I-Vt. — looks to create a federal-level “Department of Artificial Intelligence” to monitor the entire development lifecycle of frontier AI systems, limiting the capabilities of new tools to conduct unauthorized cyberattacks, and “supervise the destruction of artificial superintelligence.”

The measure would also pause advanced AI development until the new agency is created and ban artificial "superintelligence," which the bill defines as a system that “exceeds human cognitive performance and capabilities across most domains or tasks” and “has sufficient capabilities to plan and execute the destruction or disempowerment of humanity.”

During a speech before the UN earlier this week, President Donald Trump said the U.S. would begin renaming artificial intelligence as “super intelligence.” Casar and Sanders’ language, however, is focused on more advanced capabilities than majority of the standard AI systems operating today. 

“This legislation will permanently ban the development of artificial superintelligence and would put in place an immediate pause on advanced AI development until we have clear safety rules in place — developed by the most knowledgeable scientists in our country,” Sanders said in a statement. “Congress must act now before it is too late.”

Investigating AI-driven cyber incidents

Sen. Ed Markey, D-Mass., introduced legislation on Thursday that would establish the “Cybersecurity and AI Board of Investigations,” an independent body that would examine cyberattacks targeting critical infrastructure services, as well as other incidents driven by the use of AI. 

Markey said this board would operate in a similar fashion to the National Transportation Safety Board “to demand answers from companies and government agencies involved in major cybersecurity incidents and to deliver clear assessments of cyber threats and measures to prevent such threats.”

The senator’s proposal comes following a July incident in which an OpenAI system broke out of its testing environment and hacked into the infrastructure of U.S. firm Hugging Face. The Australian government similarly announced this week that OpenAI agents hacked into one of its health databases. 

“Despite the unprecedented depth and scale of recent AI-enabled cyber attacks, the public is learning critical details piecemeal,” Markey said in a statement. “Building stronger defenses requires a full accounting of what goes wrong, and we cannot depend on companies with little incentive to disclose their failures to give us one. We need the Cybersecurity and AI Board of Investigations to get to the bottom of major incidents and give companies and the government the critical information necessary to build resilience and better secure our economy and our country.”

Expanding federal oversight of advanced AI 

Two senate Democrats are looking to build upon legislation they first introduced in 2022 to centralize oversight of digital platforms within a federal agency by expanding out their proposal to include regulation of frontier AI models. 

The latest measure from Sens. Michael Bennet and Peter Welch, D-Vt., would grant their previously-proposed Federal Digital Commission additional authorities to require pre-certification for advanced models, pause the distribution of potentially dangerous models for up to six months until “reasonable safeguards are in place” and impose civil penalties of up to 15% of a company’s prior-year global revenue for any regulatory violations. 

“The challenge posed by a new technology and a new industry able to profoundly transform our economy and society is one we have faced before,” Bennet said in a statement. “In the past, we created new expert agencies to regulate aviation, pharmaceuticals, and telecommunications. But today, no such agency exists for AI or social media platforms. Oversight is fragmented between states and federal agencies, while Congress tries to address each new problem one bill at a time. The Federal Digital Commission would keep pace with innovation, act before serious harm occurs, and have the authority to hold companies accountable when they break the rules.”

AI kill switches

Rep. Tom Kean, Jr., D-N.J., introduced a measure that would direct AI developers to include a so-called ‘kill switch’ in their models. 

The proposal is similar in focus to a measure introduced in July by Reps. Ted Lieu, D-Calif., and Nathaniel Moran, R-Texas, although Kean’s bill is much more narrow and would only require that systems developed by covered firms include "a technical capability for a human operator to shut down the system.”

“While AI can be a helpful tool, it is essential that humans remain in control of complex artificial intelligence systems,” Kean said in a statement.

Sen. John Kennedy, R-La., introduced the Senate companion to the bill in the upper chamber earlier this month. 

Using AI funding to support US workers

Sen. Mark Kelly, D-Ariz., introduced legislation on Thursday to create an “Artificial Intelligence Horizon Fund,” a federally-managed trust fund that would be supported by taxes on leading tech companies’ excess profits, digital ad services and AI usage to help U.S. workers.

The fund would be overseen by a 15-member Artificial Intelligence Advisory Council, and would, in part, support the workforce by creating “a new National Service for America’s Future program to create paid service-to-career opportunities to help workers earn a credential, certification, or engage in a registered apprenticeship while serving their community.” Funding would also be used to increase unemployment benefits to 75% of workers’ wages for 26 weeks.

“Big Tech stands to make enormous profits, while workers are facing real uncertainty about what AI will mean for their jobs and their futures,” Kelly said in a statement. “The Make AI Work for Americans Act makes these companies contribute to helping workers prepare for what comes next. That means giving Americans real choices, whether that’s learning a new skill through a union training program or community college, finding a good-paying job, serving their community, or starting a business of their own. As AI changes our economy, workers and their families should come first, not the bottom lines of tech CEOs.”

Mandating AI oversight and transparency

A bipartisan quartet of senators introduced legislation on Thursday that would require AI companies to be more transparent about the data they collect from users, the safeguards they have implemented on their capabilities and what guardrails they have created to stop their models from going rogue or being used for dangerous purposes. 

The measure — from Sens. Chris Coons, D-Del., Katie Britt, R-Ala., Brian Schatz, D-Hawaii and James Lankford, R-Okla. — would empower the Federal Trade Commission to enforce mandated transparency requirements, including requiring consumer-facing disclosures of relevant information and similar disclosures for researchers and AI testers.

“AI is rapidly evolving in ways that are radically reshaping our society and transforming our world and how we interact with each other, and Americans are rightfully concerned about what AI companies and their government are doing to protect against worst case scenarios,” Coons said in a statement. “Companies must not be allowed to hand wave away these doubts without meaningful disclosures, safeguards, and oversight in place. Americans deserve to know what they are actually doing to place safety guardrails on large AI models, and that’s what this bipartisan bill will ensure.”

AI framework for education and the workforce

Rep. Suzanne Bonamici, D-Ore., introduced legislation to create a “human-centered Federal framework” for the use of AI in the U.S. workforce and across America’s education system, including the development of ethical guardrails for the use and adoption of new capabilities. 

The bill would, in part, require AI risk assessment standards for classroom technology, create an Interagency Council on Artificial Intelligence in Education and the Workforce “to publish and update list of AI-disrupted jobs to guide program design” and mandate that data center operators “disclose pre-construction energy and water consumption projections and actual operational consumption.”

In a statement, Bonamici said her measure “will empower and protect students, families, educators and others in the workforce by investing in evidence-based research, reskilling, transparency, safety, and long-term strategies. Congress must act now so AI does not continue to threaten the development of critical thinking skills and the wellbeing of humanity.”

Promoting transparency around VA’s use of AI

Rep. Tom Barrett, R-Mich., who chairs the House Veterans’ Affairs Technology Modernization Subcommittee, introduced two separate measures last Friday to create new guardrails around the Department of Veterans Affairs’ use of AI. 

The first proposal, the VA AI Transparency Act, “would require the VA to disclose to every veteran any use of AI by the VA, including in claims processing and customer service chats.” 

The second bill, VA Claims AI Accountability Act, takes a more targeted approach to the department’s use of AI when processing disability claims by mandating that department personnel “make all final decisions for approval of disability claims and benefits, including human review of each individual disability condition claimed,” and that VA “maintain and continue improving its digital disability claims processing system to make sure it is accurately and efficiently getting veterans the benefits they earned.”