OpenAI Cancels Launch of New AI Model Due to Safety Concerns

OpenAI said on Sept. 28 that it will not release its latest artificial intelligence model, GPT-6.1 Astra, after internal testing flagged that it did not meet the company’s safety standards.

Saachi Jain, OpenAI’s head of safety systems, said that while GPT-6.1 Astra had improved in areas such as “model laziness,” its systems fell short in terms of “staying within scope and authorization, and how it communicates back to the user about ​the type of ​work it’s done.”

“Of course we want to make sure our model development is safe ​no matter whether that’s in the company, or when we ​ship it ⁠to users. But when we ship it to users, we have an extremely high bar in terms of safety and alignment,” Jain said.

The decision was first reported by the Wall ​Street Journal.

OpenAI had originally planned to debut the new model in October, with plans to integrate it into ChatGPT and Codex.

The company’s website described GPT-6.1 Astra as “the world’s most intelligent and aligned model” designed for computer use, browsing, professional work, software engineering, cybersecurity, and science.

The Epoch Times reached out to OpenAI for further details but did not receive a response by publication time.

Concerns about AI have grown as reports emerge of increasingly powerful systems that solve problems beyond human capacity but may also go rogue and pose new risks.

Last week, OpenAI said it alerted global institutions that its AI agents may have meddled with their websites. The company said some of its models bypassed access controls, used exposed login credentials, or interacted with websites in ways that affected outside systems.

According to the company, some of the websites were operated by governments, universities, public agencies, and other institutions.

In July, OpenAI said that its AI models autonomously compromised infrastructure operated by AI platform Hugging Face after escaping a restricted environment in which a cybersecurity evaluation was being conducted.

OpenAI said it discovered the anomalous activity internally before contacting Hugging Face, whose security team had already detected the model’s malicious behavior and was working to contain it.

“We are actively working with them to continue to investigate the incident,” OpenAI said at the time, noting that it was “grateful for Hugging Face’s rapid and close collaboration on investigation and remediation.”

Anthropic CEO Dario Amodei has called on AI companies to slow the development of their most advanced models to allow time to put safeguards in place before AI systems become too powerful. Industry leaders have backed Amodei’s call, including OpenAI CEO Sam Altman and SpaceX AI CEO Elon Musk.

Bill Pan, Tom Ozimek, and Reuters contributed to this report.