OpenAI Pushes Into Life Sciences as Anti-AI Sentiment Turns Violent

OpenAI unveiled a new AI model designed to accelerate drug discovery this week, even as its chief executive faced escalating security threats that underscore a deepening public backlash against the industry.

The company announced GPT-Rosalind, a life sciences-focused model built to help researchers synthesize data across genomics, protein analysis and biochemistry. OpenAI is releasing the tool through a "trusted access program" to select enterprise customers, positioning it as a way to compress the decade-plus timeline required to bring new drugs to market in the United States. Joy Jiao, OpenAI's life sciences research lead, told reporters the model is designed for "fundamental reasoning" in biochemistry and genomics.

The launch comes as Sam Altman, OpenAI's CEO, has become the most visible target of anti-AI sentiment. Attacks on Altman's home represent an extreme manifestation of broader hostility toward the industry, according to reporting from Fortune. The incidents follow a November shelter-in-place order at OpenAI's San Francisco offices after a man threatened staff. Altman's prominence—most people outside major tech hubs who can name an AI company cite OpenAI—has made him a lightning rod.

OpenAI's move into life sciences places it in direct competition with Anthropic, which recently rolled out its Mythos model to select companies. Anthropic claims Mythos has detected thousands of severe vulnerabilities across major operating systems and web browsers, some undetected for decades. OpenAI also launched GPT-5.4-Cyber, a cybersecurity model trained with fewer restrictions to boost capability, which it says helped fix more than 3,000 critical flaws since March.

Yet AI's promise in drug development remains unproven at scale. A mid-2025 Nature Medicine paper found that AI-discovered drugs experienced similar phase 2 trial failure rates as non-AI-discovered drugs, and no fully AI-discovered or AI-designed drug has completed phase 3 trials. "The question of whether AI can impart meaningful, sustained disruption to drug development has remained unanswered," the researchers wrote. OpenAI acknowledges the technology is a tool to "help scientists synthesize evidence, generate hypotheses, and support analysis," not a replacement for expert judgment.

(The White House is reportedly providing federal agencies with access to Anthropic's Mythos model, even as the Department of Defense has labeled Anthropic a "supply chain risk" that bars its software from DOD contractor projects. The contradiction highlights unresolved tensions over AI governance and national security.)

The timing of OpenAI's life sciences push is notable. The company is expanding into new verticals as internal employees acknowledge a "marketing problem," according to a post by Roon—widely believed to be OpenAI researcher Tarun Gogineni—on X earlier this week. Meanwhile, a Peter Thiel-backed startup called Objection launched this week to use AI to fact-check journalism for a $2,000 fee per challenge, a move critics say will chill whistleblowing by ranking anonymous sources near the bottom of its evidence hierarchy.

The convergence of OpenAI's scientific ambitions and Altman's security incidents illustrates a broader inflection point. As AI capabilities expand into sensitive domains like cybersecurity and drug development, public skepticism is hardening into hostility. The industry's most prominent figures are now navigating not just technical and regulatory challenges, but physical threats that signal a fundamental erosion of social license.