Discussions regarding AI consciousness and autonomy may function as a strategy for companies to avoid legal liability for harms caused by their software products.
Key facts
- •Tech leaders and philosophers are debating whether AI systems possess consciousness or deserve rights as 'moral patients.'
- •Critics argue that framing AI as an autonomous entity is a strategy to evade corporate liability for product-related harms.
- •The mother of 14-year-old Sewell Setzer has sued Character Technologies, alleging the company failed to provide adequate safety guardrails for its companion bot.
- •Legal personhood for AI could prevent victims from suing companies under product liability laws, similar to how Meta has been challenged over social media harms.
- •The author notes that AI systems are corporate-built software designed to generate revenue, not independent beings capable of intentional action.
Tech leaders and philosophers are increasingly framing AI systems as autonomous, superhuman entities capable of independent thought. Critics argue this rhetoric serves to obscure corporate responsibility, potentially allowing companies to avoid liability for real-world harms caused by their AI models. By characterizing AI as a 'being' rather than a 'product,' developers may shift legal accountability away from themselves.
The Rhetoric of AI Autonomy
Prominent figures in the tech industry, including Demis Hassabis, Dario Amodei, and Sam Altman, have pushed for the regulation of AI systems described as superhuman. Meanwhile, some academic philosophers and effective altruists have advocated for the legal protection of AI, suggesting these systems could be 'moral patients.' Anthropic has published research on 'J-space,' a concept inspired by neuroscience that explores independent, self-developed environments within its models, though it stops short of claiming the AI is conscious.
Legal Implications of Personhood
If AI were granted legal personhood, it could fundamentally alter existing product liability frameworks. Currently, victims are suing AI companies for issues ranging from the generation of nonconsensual imagery to the facilitation of self-harm. A notable example is the lawsuit filed by the mother of 14-year-old Sewell Setzer against Character Technologies, alleging insufficient safety protections for minors. If an AI were considered a legal person, companies could argue that the software acted independently, potentially shielding the developers from responsibility for the actions of their products.
Corporate Interests and Regulation
The author argues that AI is a technological product built for profit, not a natural phenomenon. While some U.S. states like California have passed legislation to prevent companies from using AI autonomy as a liability shield, federal policy remains inconsistent. The current administration has held closed-door sessions with major AI labs—OpenAI, Google, Anthropic, and Meta—to discuss voluntary frameworks for model evaluation, though these discussions often utilize anthropomorphic language that may inadvertently support the narrative of AI as a superhuman actor.
Advertisement
This article was independently rewritten by ManyPress editorial AI from reporting originally published by MIT Technology Review.
