AI‑industry trust signals spike

Trust in the artificial intelligence industry is now being tested as much by politics and public reaction as by the technology itself. OpenAI spent the past week answering questions about liability, safety and security after a bill it supports in Illinois and an attack on Sam Altman’s San Francisco home drew fresh scrutiny. (openai.com) (legiscan.com) (futurism.com) The Illinois measure, Senate Bill 3444, was referred to committee on February 4, 2026. Its draft says a developer of a “frontier artificial intelligence model” would not be liable for “critical harms” if it did not act intentionally or recklessly and if it posted a safety protocol and transparency report. (legiscan.com) The same draft says companies could also satisfy the law by agreeing to European Union safety rules or by entering an agreement with a federal agency, and it would stop applying if Washington later adopted overlapping federal rules. OpenAI said on March 13, 2025 that it wanted national policy to avoid what it called “overly burdensome state laws.” (legiscan.com) (openai.com) That is the core trust problem now facing the biggest artificial intelligence firms: they say advanced systems can create severe harms, but they are also asking lawmakers to narrow liability and centralize rulemaking. OpenAI spokesperson Jamie Radice told Wired, as quoted by Futurism on April 12, that the company supports approaches that reduce serious harm while avoiding a state-by-state patchwork. (futurism.com) The public side of that debate turned physical on April 10, when San Francisco police arrested a suspect after an incendiary device was thrown near Altman’s home at about 3:45 a.m. Pacific time and threats were later made outside OpenAI’s Mission Bay headquarters. OpenAI said no one was hurt and employees should expect increased police and security presence. (futurism.com) The San Francisco Standard reported on April 12 that the attack made OpenAI’s push for a new regulatory bargain feel more urgent inside the company. Its homepage also reported a second incident at Altman’s Russian Hill home, saying early Sunday a car appeared to fire a gun at the residence and that two suspects were arrested. (sfstandard.com) This is landing during a rough stretch for OpenAI’s reputation. Futurism reported on April 12 that the company had spent 2026 dealing with backlash over a late-February Department of Defense contract, the shutdown of its Sora video app, and pressure over spending as it pursues a possible initial public offering later this year at a valuation it said could reach $1 trillion. (futurism.com) OpenAI’s position is that stronger federal coordination would protect both innovation and national security. In its March 2025 policy paper, the company said the federal government should work with the private sector and “neutralize” advantages that China could gain if United States companies had to comply with many separate state laws. (openai.com) Critics are making the opposite case: if companies warn that their systems could enable mass casualties, chemical weapons or other catastrophic misuse, they should not get broader legal shields before those risks are resolved. The Illinois draft itself uses the phrase “critical harms,” and Futurism reported that critics cited thresholds including mass deaths, injuries to more than 100 people, or more than $1 billion in property damage. (legiscan.com) (futurism.com) For now, the story is less about a single bill or a single attack than about a new phase for artificial intelligence companies: they are no longer being judged only on product launches. They are being judged on whether their safety warnings, lobbying and public accountability line up. (openai.com) (legiscan.com) (futurism.com)