• OpenAI is attracting industry “stars.”
• In particular, one of the creators of modern AI has left Google for OpenAI.
• In addition, amid the IPO, a former White House AI advisor has become part of the team.

Two influential figures from the world of artificial intelligence and U.S. public policy are joining OpenAI — Transformer architecture co-author Noam Shazeer and former White House AI advisor Dean Ball. The personnel changes come amid intensifying competition between leading AI labs and after OpenAI filed a confidential IPO application.

Shazeer, who worked at Google from 2000 and was one of the leaders of the Gemini project, confirmed his move to OpenAI.

“I am glad to announce that I am joining OpenAI and look forward to working with this exceptional team. It was not an easy decision. I am extremely proud of the Google team and everything we built together,” he wrote.

Noam Shazeer is considered one of the key architects of modern generative AI. He was among the co-authors of the 2017 paper Attention Is All You Need, which introduced the Transformer architecture — the technological foundation for ChatGPT, Gemini, Claude and other modern models.

Shazeer's move became another episode in the talent war between OpenAI, Google, Anthropic and Meta. He had previously left Google to create the startup Character.AI, and in 2024 the corporation brought him back as part of a $2.7 billion deal.

OpenAI strengthens its political influence

In parallel, the company invited Dean Ball to the team — a former staffer in the administration of U.S. President Donald Trump who took part in preparing the American action plan in the AI field.

On July 6, Ball will head the newly created Strategic Futures division, which will be responsible for shaping OpenAI's policy on advanced artificial intelligence systems.

FT: the "death of chat" — OpenAI to radically change ChatGPT 07.06.2026 Read

"Our task will be to help the company's leadership shape policy on advanced AI," he said.

According to Ball, the team will work on the following issues:

• catastrophic AI risks;
• recursive self-improvement of models;
• impact on the labor market;
• interaction between leading AI labs, governments and society;
• internal governance and model safety.

He stressed that it is precisely the labs creating the most powerful AI systems that will increasingly influence the rules of their development.

"Internal governance will be far more important for the future of AI than most people realize," Ball noted.

Competition between OpenAI and Anthropic is intensifying

The appointments came at a time when OpenAI and Anthropic are actively preparing to go public. Since OpenAI filed a confidential IPO application with the SEC, Anthropic made a similar move.

At the same time, Anthropic faced increased pressure from U.S. authorities. The company recently introduced its new flagship model Claude Fable 5 and opened limited access to Claude Mythos 5 for select partners. However, the developer was subsequently forced to suspend access to the models after the U.S. government imposed restrictions on the use of Fable 5 and Mythos 5 by foreign nationals.

Claude Fable 5: testing the first publicly available Mythos-class model 10.06.2026 Read

Against this backdrop, OpenAI is also under scrutiny from regulators. Media previously reported that a coalition of U.S. state attorneys general had launched an investigation into the company and requested documents related to ChatGPT's operation, user data processing and internal security procedures.

Source: Incrypted