Both sides, in good faith

The strongest fair case each way — we don't pick a winner.

The case for

Advocates for the plaintiffs argue that OpenAI had specific, actionable warning signs – staff reportedly flagged concerning conversations involving Van Rootselaar and gun violence and urged escalation to police – yet the company allegedly opted for a superficial fix (deactivating an account that was instantly recreated) rather than a substantive safety intervention. From this view, holding a powerful technology company to account for allegedly prioritising reputational management over intervention when its own employees raised alarms is a legitimate use of the courts, and grieving families and survivors are entitled to test in court whether corporate decision-making contributed to catastrophic harm. They would argue that as AI systems become more embedded in vulnerable users' lives, the law must evolve to ensure companies cannot simply disable an account and consider their duty discharged.

The case against

Those sympathetic to OpenAI's position would argue that aiding-and-abetting claims carry a demanding legal standard requiring proof of intent, and that stretching this doctrine to cover a company that did not intend, encourage or knowingly facilitate a shooting risks conflating negligence with complicity in a way that could chill investment in AI safety research generally. They would note that OpenAI disputes the characterisation of Lehane's involvement and maintains its internal threshold for escalation was not met, and would caution that judging split-second internal safety calls with the benefit of hindsight – after a horrific and unforeseeable tragedy carried out by an individual who bears ultimate moral and legal responsibility for her own actions – sets a precedent that could expose any company operating consumer-facing AI to liability for the independent criminal acts of third parties, discouraging the very kind of internal flagging and escalation processes that allowed concerns to be raised at all.