OpenAI, the company behind ChatGPT, is cutting access to its new cybersecurity AI tool, GPT -5.5 Cyber. The move came shortly after its CEO, Sam Altman, criticized Anthropic, the firm behind Claude AI, for doing the same. The latest limitation shows growing tension in the AI space as firms try to launch new ideas amid AI safety concerns.
Claim 55% Off TipRanks
Forget margin or options. Here's how the pros trade NVDAOpenAI Limits Cyber Tool after Criticizing Anthropic
Altman said in an X post earlier this week that GPT -5.5 Cyber will be rolled out in a few days. However, it will be accessed only by a limited group of users, whom he described as “critical cyber defenders.” These users will apply for access through a vetting process, in which they submit personal details and their intended use of the AI before gaining access.
The Cyber AI tool is said to be designed for advanced cybersecurity tasks, such as testing systems for flaws, finding weak points, exploiting software bugs, and studying malware. While these features can help companies protect themselves, they could also be misused by bad actors for their own gain.
Notably, Cyber’s restricted access comes after Altman threw a jab at Anthropic, its biggest AI rival, for limiting access to Mythos. He called Anthropic’s approach a “fear-based marketing” tactic. He also said the firm was exaggerating risks to keep control of its technology. Ironically, OpenAI is now taking the same path by curbing access to its own AI tool while it tries to expand its use to a controlled group.
AI Rivalry Increases as Safety Concerns Grow
Anthropic had earlier released its new Mythos AI model to a small group of enterprise users. The firm said it had limited access to the tool because it was too powerful for public release, given the risk of misuse by cybercriminals. This sparked various reactions from critics, as some supported the cautious approach while others felt the warnings were overstated.
Despite Anthropic’s restrictions, reports showed that an unauthorized group had gained access to the tool last week. This raised questions about how strong the company’s control and security over the AI tool really is.
Meanwhile, OpenAI is working with the U.S. government to expand access to Cyber in a more controlled way. This ensures that the system is used only by people with verified cybersecurity expertise.
Additionally, it shows that firms in the AI industry are not just developing powerful tools but are becoming more careful about who gets to use them. Although Anthropic and OpenAI’s rivalry continues, they are both taking similar steps to ensure the safety of their new AI tools.
Which AI Stock Can I Buy Now?
Wall Street analysts rate Nvidia (NVDA), Meta Platforms (META), Microsoft (MSFT), and Alphabet Class A (GOOGL) as Strong Buy, based on TipRanks consensus data. Among these stocks, NVDA has the highest upside potential 37.49% and an average price target of $274.38. For more information on performance, ratings, and price targets of each of these stocks, visit TipRanks Stocks Comparison Center.
