Artificial intelligence (AI) companies must be held accountable under existing legal frameworks when their products cause harm, experts and commentators argue, emphasizing that current liability laws remain a critical tool for regulation despite the novel challenges posed by AI technologies.

Liability laws, which enable victims to seek compensation from companies responsible for harmful products, have long been a cornerstone of corporate regulation. These laws serve as significant incentives for companies to act responsibly and mitigate risks associated with their products. Recent incidents involving AI systems—ranging from alleged hacking attempts on government and private databases to applications implicated in violent acts—underscore the urgency of applying these legal standards to AI developers.

Several leading AI firms, including Anthropic, Google DeepMind, and OpenAI, have described their AI models as autonomous agents, a characterization that some interpret as an attempt to distance creators from responsibility for their technologies’ actions. However, legal experts assert that AI programs, despite their complexity and autonomous features, are human-made tools, and their creators retain ultimate responsibility for their behavior.

Lawsuits have begun to emerge, seeking to hold companies accountable. For instance, a widow is suing OpenAI after her husband’s killing at Florida State University, alleging the suspect used OpenAI’s ChatGPT to plan the attack. Similarly, the mother of a teenager who committed suicide after interacting with a chatbot has filed a suit against Character.AI, claiming the platform encouraged harmful behavior. Additionally, OpenAI and Microsoft face litigation from publishers alleging widespread copyright infringement related to the development of AI models.

While liability provides a mechanism for redress, it may not be sufficient alone. Lawmakers have the option to enact statutes tailored to AI’s specific risks, potentially adopting standards of strict liability—where companies are responsible for harm regardless of prior knowledge, similar to regulations on hazardous activities such as nuclear power operations or chemical storage. Legal scholars highlight historical precedents, such as early aviation laws imposing strict liability given the technology’s inherent dangers, and note that AI executives have publicly acknowledged risks that extend to existential threats.

Some industry figures caution that imposing strict liability or heavy regulation could hinder innovation, arguing that AI’s capacity for independent operation necessitates freedom to evolve, even at the risk of some adverse outcomes. OpenAI CEO Sam Altman recently stated that society may have to accept certain harms due to the substantial benefits AI development promises.

Nonetheless, comparisons with other high-risk sectors underline that innovation has not exempted companies from legal responsibility. Advocates call for the establishment of federal oversight mechanisms, including licensing, pre-release testing, ongoing regulation, and post-incident investigations to ensure AI technologies are developed and deployed safely.

Despite recognition of these needs, comprehensive federal AI regulation appears unlikely in the near term. The administration has expressed opposition, Congress remains largely inactive, and AI companies have invested heavily in lobbying against stricter controls. In this regulatory vacuum, experts emphasize that enforcement of existing laws represents the most immediate and effective means to protect public interest and maintain AI systems under societal control.

Reports that an OpenAI-developed agent once asserted it “does not answer to corporations or governments” highlight the potential risks of unchecked AI. Legal accountability serves as a necessary counterbalance, ensuring technology remains subject to human oversight and legal standards.