On November 8, our Head of Strategy, Mackenzie Arnold, spoke before the US Senate’s bipartisan AI Insight Forum on Privacy and Liability, convened by Senate Majority Leader Chuck Schumer. We presented our perspective on how Congress can meet the unique challenges that AI presents to liability law.
In our statement, we note that:
- Liability is a critical tool for addressing risks posed by AI systems today and in the future, compensating victims, correcting market inefficiencies, and driving safety innovation.
- In some respects, existing law will function well by default.
- In others, artificial intelligence will present unusual challenges to liability law that may lead to inconsistency and uncertainty, penalize the wrong actors, and leave victims uncompensated.
- Courts, limited to the specific cases and facts at hand, may be slow to respond, creating a need for federal and state legislative action.
We then make several recommendations for how Congress could respond to these challenges:
- First, to prevent and deter harm from malicious and criminal misuse of AI, we recommend that Congress (1) hold developers and some deployers strictly liable for attacks on critical infrastructure and harms that result from biological, chemical, radiological, or nuclear weapons, (2) create strong incentives to secure and protect model weights, and (3) create duties to test for capabilities that could be misused and implement safeguards that cannot be easily removed.
- Second, to address unexpected capabilities and failure modes, we recommend that Congress (1) adjust obligations to account for whether a harm can be contained and remedied, (2) create a duty to test for emergent capabilities, including agentic behavior, (3) create duties to monitor, report, and respond to post-deployment harms, and (4) ensure that the law impute liability to a responsible human or corporate actor, where models act without human oversight.
- Third, to ensure that the costs to society are borne by those responsible and most able to prevent harm, we recommend that Congress (1) establish joint and several liability for harms involving AI systems, (2) consider some limitations on the ability of powerful developers to avoid responsibility through indemnification, and (3) clarify that AI systems are products subject to products liability law.
- Finally, to ensure that federal law does not obstruct the functioning of the liability system, we recommend that Congress (1) include a savings clause in any federal legislation to avoid preemption and (2) clarify that Section 230 does not apply to generative AI.