OpenAI targets specialised AI

OpenAI is carving its chatbot business into industry tools, starting with a life-sciences model and a cyber model aimed at vetted defenders. (openai.com) On April 16, OpenAI introduced GPT‑Rosalind, which it said is built for biology, drug discovery and translational medicine rather than general chat. The company said the model is optimized for chemistry, protein engineering and genomics workflows. (openai.com) Drug discovery is the long front end of making a medicine: researchers pick a biological target, test ideas in the lab, and narrow them before human trials begin. OpenAI said that process typically takes about 10 to 15 years from target discovery to U.S. regulatory approval, and that earlier gains can improve later experiments. (openai.com) OpenAI is also widening access to GPT‑5.4‑Cyber, a version of GPT‑5.4 tuned for defensive security work such as finding weaknesses and helping responders move faster. The expansion, announced April 14, opens its Trusted Access for Cyber program to thousands of verified individual defenders and hundreds of teams protecting critical software. (openai.com) The cyber push comes with named partners and government evaluators. OpenAI said Bank of America, Cisco, Cloudflare, CrowdStrike, Goldman Sachs, Oracle, Palo Alto Networks and Zscaler are among the organizations already participating, and that the U.S. Center for AI Standards and Innovation and the U.K. AI Security Institute are evaluating the model’s cyber capabilities and safeguards. (openai.com) OpenAI has been moving its main GPT line toward professional work at the same time. When it released GPT‑5.4 on March 5, the company said the model was designed for professional tasks, offered up to 1 million tokens of context, and was available in ChatGPT, the application programming interface and Codex. (openai.com) The company is pairing those specialized models with product controls in ChatGPT itself. In release notes published April 16, OpenAI said ads are being rolled out on Free and Go plans in Australia, New Zealand and Canada, but not in chats about health, mental health or politics, and not for users it knows or predicts are under 18. (help.openai.com) A separate OpenAI help page says advertisers in dating, health, financial services and politics are excluded from ChatGPT ads for now, and that Plus, Pro, Business, Enterprise and Edu plans do not carry ads. The same page says ads are shown below responses, labeled as sponsored, and do not influence answers. (help.openai.com) Security specialists described the cyber release as a broader-access alternative to more tightly restricted models from rivals. Security Magazine reported April 16 that OpenAI is scaling access to thousands of verified defenders, while Anthropic limited Claude Mythos to a smaller set of partners. (securitymagazine.com) Taken together, the April releases show OpenAI selling less of a single chatbot and more of a menu: one model for lab researchers, one for security teams, and a consumer product with tighter controls around ads and sensitive topics. (openai.com, openai.com, help.openai.com)