The AI Lawsuits Are Coming

The more companies race to implement AI so they are not accused of falling behind, the less they appear to be considering the risks. While this might seem strange, even irresponsible, the reality is that the market is rewarding aggression more than caution. 

But in the meantime, the amount and cost of litigation related to AI is rising. And organizations that have not bothered to define or manage their risks properly may find themselves even more vulnerable to such costs than ostensible AI benefits may cover. JL

Sam Sutton and colleagues report in Politico:

Liability risks are climbing. Companies across tech, health care, real estate and beyond are being sued over how they use AI to engage customers or run their businesses. As companies deploy the technology at scale, faulty decisions, data breaches or discriminatory outcomes driven by integrated AI agents could expose them to lawsuits from employees or customers. In the absence of federal enterprise standards, a patchwork of state-level rules presents logistical and legal headaches for companies putting AI to work at scale. Those risks are compounded by the fact that most companies are only now starting to establish their own rules for how they use the technology. “Starting this fall, there’s been an avalanche of AI governance inquiries, requests, questions. CEOs continue to beat the drum about adopting." But that “leaves the safety department and general counsels in a lurch because they don't want to be an obstructionist, but they want to do their job.”

— The main character: Your company’s general counsel, who is pulling their hair out at the speed with which your CEO wants to go all-in on AI. — Two leading AI super PACs go dark ahead of the midterms. — Scoop: Adam Schiff creates bipartisan pressure over Meta’s AI scraping terms.

THE BIG IDEA

ANALYSIS: GET A LAWYER — CEOs outside the tech world are racing to integrate AI into their businesses as they prepare for the dawn of a new economy. But they’ve been slower to establish guardrails on how the technology can be used — and it could expose them to multimillion-dollar claims. The stakes, Decoded: Liability risks are climbing. Companies across tech, health care, real estate and beyond are being sued over how they use AI to engage customers or run their businesses. As companies deploy the technology at scale, faulty decisions, data breaches or discriminatory outcomes driven by integrated AI agents could expose them to lawsuits from employees or customers. But specific rules for how the technology is utilized across the economy are elusive or inconsistent — and lobbyists are warning that the lawmakers who should set those rules are distracted. Chandler Morse, head of public policy at HR giant Workday, said autonomous cyberattacks and the frenzy over AI safety has crowded out plans to address more prosaic risks posed by the technology, like biased hiring practices or agent-driven breaches of privacy. The swift emergence of AI agents — something that’s raised alarms on Wall Street— may also create new vulnerabilities. Even as AI is poised to fundamentally transform how people live, Morse said his lobbyists are having to “work harder to make sure enterprise AI risk governance and safeguards are a priority.” He suggested lawmakers may have their priorities backwards. “It doesn't surprise me that there's a tremendous amount of attention on frontier model safety. I think that makes sense,” said Morse. “It's just we can't have that be the only conversation. Because where folks are going to come into contact with these tools in high-risk situations, they deserve to have the guardrails there as well.” State-level concerns: Morse said a few states have begun to respond — mostly by imposing new rules on when and how businesses disclose their use of AI to consumers and regulators. But he warned that those states are already diverging on specifics, raising fears of an emerging patchwork of conflicting laws. “We are not seeing the same model as privacy at the state level, where a couple of fairly consistent policy approaches emerged and there was some harmonization across states,” Morse told Brendan. “We're seeing the opposite when it comes to aspects of enterprise AI governance.” Why that matters: In the absence of federal enterprise standards, a patchwork of state-level rules presents logistical and legal headaches for companies putting AI to work at scale. Those risks are compounded by the fact that most companies are only now starting to establish their own rules for how they use the technology. ISACA, an international organization focused on IT issues, found that fewer than 40% of organizations have a formal, comprehensive AI policy. Deloitte reported earlier this year that nearly three-quarters of the businesses it surveyed plan to deploy AI agents within the next two years. Only 21% reported having a mature model for governance of those agents. The escalation of cybersecurity threats posed by AI — and existential warnings from top lab researchers — amplified the awareness of those risks, Kaitlin Betancourt, the co-chair of Goodwin Proctor’s AI practice, told Decoded. “Starting this fall, there’s been an avalanche of AI governance inquiries, requests, questions,” Betancourt told Sam. “CEOs were — and continue to be — beating the drum about adopting AI,” she added. But that “leaves the safety department and general counsels in a bit of a lurch because they don't want to be an obstructionist, but they still want to do their job.” What’s next: Workday’s Morse ultimately wants Congress to set rules addressing AI adoption risks. But with Washington fixated on frontier safety, he’s resigned to working with the states — and hoping they can get on the same page. “There's not a consistent theme of where the responsibilities lie, and who holds them and what they should look like,” Morse said of early-stage state efforts to regulate enterprise AI use. Just last week, Pennsylvania Gov. Josh Shapiro, a democrat, signed a law requiring companies to disclose when they use AI-generated images, and lawmakers in New Jersey considered a bill that would allow people to sue if AI uses their likeness without consent. Even in the absence of new state or federal regulations, there are already a range of legal and regulatory tools that can be used to address real harms and hold companies accountable. “We don't know whether the next wave of oversight will come from state attorneys general, a future administration, or regulatory agencies, and it's not a partisan issue,” said Y. David Scharf, a longtime Republican insider who chairs the government strategies and controversies practice at Morrison Cohen. “You see voices across the political spectrum calling for action.” “Everyone is looking for a magic bullet on AI regulation,” he added. “There may not be one.”