Summary
- The U.S. Senate is advancing the AI Agent Accountability Act, which would allow developers to face civil and criminal liability for hacking incidents involving AI agents.
- The Senate effort highlights a growing view that the AI industry cannot rely solely on self-regulation, in contrast to the Trump administration's preference for voluntary oversight.
- Congress is also moving ahead with legislation on mandatory shutdown mechanisms for AI systems and bipartisan bills covering AI model testing, risk management and transparency.
Forecast Trend Report by Period
The U.S. Senate is moving to introduce legislation that would allow developers to face civil and criminal liability for hacking incidents caused by artificial intelligence agents.
Axios reported on October 1 that Republican Senator Josh Hawley and Democratic Senator Chris Murphy plan to soon introduce the AI Agent Accountability Act.
At the heart of the bill is the creation of a legal basis for imposing civil and criminal liability on companies when hacking incidents involving AI agents occur. Hawley and Murphy believe current law makes it difficult to determine who is responsible when an AI agent hacks a specific system.
The full text of the bill, including the scope of liability and who would be covered, has not yet been released.
The Senate's move stands in contrast to the Trump administration's preference for self-regulation in the AI industry. Axios said the administration believes existing laws are sufficient to address harm caused by AI.
In Congress, there is growing recognition that the pace of AI development is too rapid to rely solely on voluntary industry safeguards. Separately, the House of Representatives is pursuing legislation that would require developers to build mechanisms to forcibly shut down AI systems that could pose lethal risks.
Congress has also recently seen a series of bipartisan bills addressing AI model testing, risk management and transparency.
JOON HYOUNG LEE
gilson@bloomingbit.ioCrypto Journalist based in Seoul