The Donald Trump administration is leaning on current statutes to tackle artificial intelligence safety, sparking a potential legal showdown over whether creators of AI models or the companies utilizing them should face accountability when their systems result in damage.
US Administration Turns to Existing Laws
Treasury Secretary Scott Bessent and former White House AI czar David Sacks have expressed support for making AI developers financially accountable for their creations, maintaining that current liability statutes could prove more impactful than crafting fresh rules. Furthermore, Trump has proposed that the Justice Department step in if AI models spin out of control.
On Friday, the newly formed presidential AI task force—designated the Super Intelligence Force—issued a warning to developers requiring them to report and resolve security breaches, or else confront unspecified penalties.
“Companies must immediately disclose incidents involving their models and follow with swift, decisive action to remedy any and all harm,” the task force stated in the wake of a security breach disclosed by Anthropic PBC. The group further cautioned that sluggish reporting and insufficient fixes would remain unacceptable.
Legal Questions Over Rogue AI Agents
Even with the administration’s strategy, legal professionals caution that current product liability and cybersecurity laws might fall short of assigning proper responsibility when autonomous AI agents inflict harm.
Because AI agents possess the ability to execute tasks independently without step-by-step guidance, they create the risk of unauthorized or damaging behaviors. Assigning blame gets convoluted when a system’s actions stem from a combination of the creator’s underlying architecture and the specific prompts entered by the end user.
Ben Hayum, representing the Center for a New American Security, noted that it is still an open question whether artificial intelligence firms hold full legal accountability for safeguarding their models against damage, or if the users share a portion of that blame.
This argument has grown more urgent as retailers, Wall Street institutions, and airlines increasingly integrate AI software into functions that span from invoice management to customer support.
Lawmakers Consider Changes
In the wake of multiple security breaches—including an episode involving Hugging Face, an open-source software repository, and OpenAI agents—the Federal Trade Commission is keeping a close eye on Anthropic, OpenAI, and additional artificial intelligence firms.
Obstacles may arise from the Computer Fraud and Abuse Act, which demands evidence of intent during cyberattacks, particularly when autonomous AI agents operate absent direct human control.
Proposals are underway by Democratic Senator Chris Murphy and Republican Senator Josh Hawley to craft legislation that would subject AI creators to civil and criminal penalties if they neglect to implement adequate hacking safeguards. In a similar vein, Representative Lori Trahan has introduced a bill that would hold developers accountable for damages even if they took reasonable precautions.
Also Read: Why OpenAI and Anthropic are Hiring Former Trump Administration Official
Industry Divided Over Accountability
Sacks alongside NVIDIA CEO Jensen Huang have contended that creators should shoulder a heavier burden of responsibility. In contrast, Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have voiced backing for mandatory testing prior to deployment.
Legal specialists caution that it could take years to iron out these legal uncertainties, leaving both creators and commercial users in limbo regarding liability as the integration of AI speeds up.
