In recent weeks, fears over the potential of AI tools in terms of cybersecurity have been making the rounds globally. The US government has imposed restrictions on AI models like Anthropic’s Claude Mythos 5 and Fable 5 for such concerns. It is believed that AI tools can be exploited by bad actors to potentially hack into systems. And now, Apple has released new security patches, earlier than its initial plan, amid such concerns.
Apple has released iOS 26.5.2, iPadOS 26.5.2 and macOS Tahoe 26.5.2. The new updates bring a set of security fixes that had originally been meant for version 26.6 of each operating system.
As per a report from Reuters, Apple said that it was adapting to the reality that AI could shorten the time attackers need to exploit known flaws. The company stated there was no evidence that any of the newly patched vulnerabilities had been used in attacks. However, the Cupertino giant added that the gap between making fixes public and placing them in customers' hands needed to be reduced.
The move marks a change from Apple's usual software release pattern. Unless security researchers uncover an active hacking campaign involving a previously unknown flaw, the company generally includes security fixes in the move from one iOS version to the next, such as from 26.5 to 26.6. In the meantime, developers and testers use beta releases of the upcoming update to identify and fix problems before the wider rollout.
Apple added that in the security notes that the updates contained fixes first made available in the beta versions of iOS 26.6, iPadOS 26.6 and macOS Tahoe 26.6, hinting that the company had decided to move them into the public release sooner than expected.
Keep in mind that in recent weeks, Anthropic’s Mythos Preview was able to spot thousands of bugs in some of the most widely used software globally as part of Project Glasswing. Apple is part of Project Glasswing, allowing the company to use the advanced AI model to spot potential vulnerabilities across its systems. It is not known whether Mythos played any role in Apple's decision to release the fixes ahead of schedule.
This is not to say that only US AI labs are making models that can pose a concern in cybersecurity. Similar capabilities are also emerging outside the US. Tokyo-based Sakana AI has said its Fugu system can rival Anthropic's models on several benchmarks, while China's 360 Security Technology has introduced Tulongfeng, which it says can compete directly with Mythos. Z.ai has also hinted that it can have a Mythos-level model in the future after releasing the GLM-5.2 model.
In recent weeks, fears over the potential of AI tools in terms of cybersecurity have been making the rounds globally. The US government has imposed restrictions on AI models like Anthropic’s Claude Mythos 5 and Fable 5 for such concerns. It is believed that AI tools can be exploited by bad actors to potentially hack into systems. And now, Apple has released new security patches, earlier than its initial plan, amid such concerns.
Apple has released iOS 26.5.2, iPadOS 26.5.2 and macOS Tahoe 26.5.2. The new updates bring a set of security fixes that had originally been meant for version 26.6 of each operating system.
As per a report from Reuters, Apple said that it was adapting to the reality that AI could shorten the time attackers need to exploit known flaws. The company stated there was no evidence that any of the newly patched vulnerabilities had been used in attacks. However, the Cupertino giant added that the gap between making fixes public and placing them in customers' hands needed to be reduced.
The move marks a change from Apple's usual software release pattern. Unless security researchers uncover an active hacking campaign involving a previously unknown flaw, the company generally includes security fixes in the move from one iOS version to the next, such as from 26.5 to 26.6. In the meantime, developers and testers use beta releases of the upcoming update to identify and fix problems before the wider rollout.
Apple added that in the security notes that the updates contained fixes first made available in the beta versions of iOS 26.6, iPadOS 26.6 and macOS Tahoe 26.6, hinting that the company had decided to move them into the public release sooner than expected.
Keep in mind that in recent weeks, Anthropic’s Mythos Preview was able to spot thousands of bugs in some of the most widely used software globally as part of Project Glasswing. Apple is part of Project Glasswing, allowing the company to use the advanced AI model to spot potential vulnerabilities across its systems. It is not known whether Mythos played any role in Apple's decision to release the fixes ahead of schedule.
This is not to say that only US AI labs are making models that can pose a concern in cybersecurity. Similar capabilities are also emerging outside the US. Tokyo-based Sakana AI has said its Fugu system can rival Anthropic's models on several benchmarks, while China's 360 Security Technology has introduced Tulongfeng, which it says can compete directly with Mythos. Z.ai has also hinted that it can have a Mythos-level model in the future after releasing the GLM-5.2 model.
Leave a Reply