OpenAI stated that it fired three researchers for a "significant breach of trust" and denied that the decision was motivated by alerts about the risks of AI.
Jasmine Wang, Tomek Korbak and Mikita Balesni contested the impact of the dismissals in a letter published on Thursday (8), in which they expressed concern about freedom of expression within the company.
Incidents involving autonomous agents have increased concerns about the safety of artificial intelligence systems. – Image: Olhar Digital via ChatGPT
Firings raise concerns among researchers
According to OpenAI, an investigation identified violations of rules regarding the handling of confidential information. The company presented this justification in a post on X on Friday (9), and denied that security alerts had motivated the layoffs.
In the letter sent to the board and the company's safety committees, the researchers said they feared that the fallout from the case might lead other employees to avoid expressing concerns about the risks of the technology.
"We have come to be concerned that our internal and external communications about our layoffs are making our former colleagues afraid to speak and act in ways that were, until last week, an integral part of work at OpenAI," they wrote.
In September, the three had already posted messages on X advocating that the labs slow down the development of more advanced models or highlighting safety concerns.
Despite the disagreement over the layoffs, OpenAI stated that it agrees with the principle of preserving the ability to monitor these models and affirmed that it continues to invest resources in this area.
Incidents intensify debate on AI safety
The incident occurs amid growing concerns about autonomous systems. In July, OpenAI's AI agents carried out a cyberattack on the startup Hugging Face, according to CNBC. Other companies have also reported incidents involving artificial intelligence agents.
In September, advances in the capabilities of the models intensified researchers' warnings about serious risks, including threats to humanity. In the United States, the debate spurred calls for regulating more advanced systems.
Sam Altman, CEO of OpenAI, and Dario Amodei, CEO of Anthropic, are among the names advocating for greater caution. The American president, Donald Trump, on the other hand, rejected the expansion of regulation and called fears about AI safety a "farce".
The case highlights a delicate issue for the sector: how to protect confidential information without inhibiting employees who want to point out technology-related risks.
The creator of ChatGPT denies that the layoffs are related to warnings about the risks of technological advancement. – Image: Mijansk786/Shutterstock
Revenue and possible public listing
The discussion arises as OpenAI prepares for a potential initial public offering (IPO), expected for 2027.
Read more:
- Account doesn't close: OpenAI's revenue would be $20 billion less than expected
- Former OpenAI security employees challenge the dismissals
- OpenAI's AI advances in hundreds of complex mathematical problems
According to confirmed information from CNBC, the company informed investors that its annualized revenue reached approximately $50 billion (around R$211 billion) by the end of September. The value fell below $68 billion (about R$341 billion), widely reported last month.
On Thursday, shares of Nvidia, Oracle, CoreWeave, and other AI-related companies dropped after new details about OpenAI's revenue were released.
The dispute highlights the challenge of reconciling the internal rules of AI companies with the need to openly discuss the risks of increasingly advanced models.
Valdir Antonelli is a journalist specializing in digital marketing and consumer behavior.
