OpenAI, the creator of ChatGPT, announced on Friday that it is limiting the launch of its latest artificial intelligence model following a request from the U.S. government under President Donald Trump’s administration. This move is part of a unique government review process aimed at assessing potential cybersecurity threats posed by AI products.
The new AI model, named GPT-5.6 Sol, will initially be accessible only to a select group of authorized partners approved by the Trump administration, according to OpenAI. The company emphasized that it does not intend for this government oversight to become a standard practice in the long term.
OpenAI stated that the current restricted release of the AI model is a temporary measure as it moves towards broader availability in the upcoming weeks. This controlled rollout strategy by OpenAI mirrors recent actions taken by the government against Anthropic, a competitor, which had to take down its AI models, Fable 5 and Mythos 5, shortly after their public release to comply with a directive from the Trump administration.
The White House confirmed its ongoing collaboration with leading AI research labs to address the evolving challenges associated with scaling up AI technologies. Concerns have escalated following Anthropic’s warning earlier this year about the potential for its Mythos model to identify software vulnerabilities that could be exploited by malicious actors, posing risks to global computer networks.
President Trump’s executive order on AI oversight, signed in June, established a framework requiring the federal government to assess the national security implications of advanced AI systems for up to 30 days before their public launch. OpenAI clarified that its new Sol model is primarily geared towards identifying and resolving vulnerabilities rather than executing cyberattacks, aligning with the company’s risk assessment criteria. However, the company acknowledged the possibility of unforeseen risks, particularly when the model is combined with other tools.
The decision to cautiously introduce the enhanced capabilities of the Sol model is a proactive measure by OpenAI to mitigate potential risks and ensure a phased deployment. A wide range of cybersecurity experts have criticized the government’s actions that led to the shutdown of Anthropic’s Fable model, emphasizing the lack of substantial evidence supporting the move.
Alex Stamos, a cybersecurity expert and former chief security officer at Meta, voiced skepticism about the necessity of the government’s intervention, highlighting that the perceived risks with the AI models are not unique and are present in other publicly available AI systems globally, including those from China. Stamos argued that such actions could hinder the U.S.’s competitiveness in the AI race against China.
The evolving landscape of AI governance and security highlights the complexities faced by technology companies and governments in balancing innovation with risk management in the digital age.