As a month filled with concerning news about artificial intelligence comes to an end, OpenAI – one of the leading companies in the field – said that it was hitting the breaks on the release of its new model.
This news came out in an exclusive article The Wall Street Journal published Monday. According to the piece, OpenAI said it was scrapping the expected release of GPT 6.1 Astra “over safety concerns that researchers raised during internal testing, in one of the clearest signs so far that agent misbehavior could stymie the industry’s rapid progression.”
OpenAI CEO Sam Altman and other industry leaders, including Anthropic’s Dario Amadei, have recently been issuing public warnings about the advances in AI and potential dangers. Earlier this month, Audacy reported on OpenAI’s disclosure of incidents where its AI went rogue and announced a new program for reporting cases of “misalignment” a term that refers to AI not following instructions and not remaining aligned with what humans actually want it to do.
There was also that X post from a former Anthropic employee that warned about AI potentially destroying humanity in the next decade. OpenAI has continued to release information about safety cases, including the Sept. 25 disclosure that “we have identified 53 instances to date where user-provided images were posted to image-hosting sites as links that weren’t publicly listed.”
OpenAI was planning to release the new GPT-6.1 Astra model next month, The Wall Street Journal said. Per OpenAI, the new model was “more capable than the company’s previous models in completing challenging tasks from end-to-end without human assistance, as well as writing,” but that it performed poorly on tests measuring alignment and it showed higher levels of deception than other models. It would also push ahead on tasks without asking for user permission.
In a statement provided to The Hill, Saachi Jain, OpenAI’s head of safety systems said GPT-6.1 Astra “didn’t quite meet the bar in terms of staying within scope authorization, and how it communicates back to the user about the type of work it’s done.”
Instead of releasing the model, OpenAI “will focus on improving the safety of future models, which it expects to be even more capable,” said the WSJ.
It did recently announce the release of GPT‑6.1 Sol, “an upgrade to GPT‑6 Sol that nearly matches GPT‑6 Astra’s intelligence on agentic coding, computer use, and professional work at one-fifth of Astra’s standard input and output token prices,” per the company description.
Another security incident prompted a Wednesday update from OpenAI about “disrupting a coordinated model-distillation campaign.” It blamed the company Moonshot AI for mass data extraction of its AI models, as Bloomberg reported on.
“Adversarial distillation poses safety and national security risks,” OpenAI noted in its update.
In addition to AI industry leaders, tech pioneer Bill Gates, former President Barack Obama and Pope Leo XIV have issued warnings about the developing technology. In Pope Leo’s encyclical letter on the topic of AI the pontiff said: “We cannot be satisfied with merely calling for the moralization of machines — the so-called ‘alignment’ of AI with human values – without also having the courage to insist on a further condition: the possibility of openly discussing the ethical frameworks involved and subjecting them to shared standards of social justice. Otherwise, those who control AI will impose their own moral vision, which will become the invisible infrastructure of these systems.”
On Tuesday, President Donald Trump, Speaker Mike Johnson (R-La.), and AI executives announced that they had “signed a voluntary and ‘morally binding’ agreement on AI,” per CSPAN. Trump said he saw “tremendous self-policing” among the AI leaders and repeatedly said that the group had decided to rename AI to “super intelligence” or SI. He has previously downplayed concerns about AI.
