AI cyber arms race: Google stopped hackers exploiting zero-days

Artificial intelligence is rapidly becoming one of the biggest tools in cybersecurity, for both defenders and attackers.

Artificial intelligence is rapidly becoming one of the biggest tools in cybersecurity, for both defenders and attackers.

Now, Google says it has stopped hackers from using AI to carry out what could have become a large-scale cyberattack campaign.

Now, Google says it has stopped hackers from using AI to carry out what could have become a large-scale cyberattack campaign.

Stay up to date with the latest news. Follow KT on WhatsApp Channels.

In a new report published by Google’s Threat Intelligence Group (GTIG), the company revealed that it detected cybercriminals using an AI model to identify and exploit a previously unknown software flaw, commonly referred to as a zero-day vulnerability.

In a new report published by Google’s Threat Intelligence Group (GTIG), the company revealed that it detected cybercriminals using an AI model to identify and exploit a previously unknown software flaw, commonly referred to as a zero-day vulnerability.

According to Google, the hackers planned to use the exploit in a “mass vulnerability exploitation operation” that could have targeted organisations at scale. The company said it has “high confidence” that the AI system was being used to help bypass two-factor authentication protections.

According to Google, the hackers planned to use the exploit in a “mass vulnerability exploitation operation” that could have targeted organisations at scale. The company said it has “high confidence” that the AI system was being used to help bypass two-factor authentication protections.

“The criminal threat actor planned to use it in a mass exploitation event but our proactive counter discovery may have prevented its use,” Google wrote in the report.

“The criminal threat actor planned to use it in a mass exploitation event but our proactive counter discovery may have prevented its use,” Google wrote in the report.

The company did not reveal the identity of the hacking group involved, though it clarified that its own Gemini AI model was not used in the operation.

The company did not reveal the identity of the hacking group involved, though it clarified that its own Gemini AI model was not used in the operation.

The findings add to growing concerns how AI is changing the cybersecurity landscape. While AI tools are helping companies strengthen defenses, they are also giving hackers faster ways to discover software weaknesses, automate attacks, and create more advanced malware.

The findings add to growing concerns how AI is changing the cybersecurity landscape. While AI tools are helping companies strengthen defenses, they are also giving hackers faster ways to discover software weaknesses, automate attacks, and create more advanced malware.

Google highlighted several examples of threat actors using publicly available AI tools, including OpenClaw, for vulnerability discovery and cyberattack planning. The report noted that groups linked to China and North Korea have shown “significant interest” in using AI for offensive cyber operations.

Google highlighted several examples of threat actors using publicly available AI tools, including OpenClaw, for vulnerability discovery and cyberattack planning. The report noted that groups linked to China and North Korea have shown “significant interest” in using AI for offensive cyber operations.

The warning comes amid increasing caution across the AI industry. Earlier this year, Anthropic delayed the release of its Mythos AI model over concerns it could help criminals exploit legacy software vulnerabilities. The move reportedly triggered discussions involving technology firms and the White House.

The warning comes amid increasing caution across the AI industry. Earlier this year, Anthropic delayed the release of its Mythos AI model over concerns it could help criminals exploit legacy software vulnerabilities. The move reportedly triggered discussions involving technology firms and the White House.

More recently, OpenAI also announced a limited rollout of GPT-5.5-Cyber, a cybersecurity-focused AI model designed specifically for vetted security teams.

More recently, OpenAI also announced a limited rollout of GPT-5.5-Cyber, a cybersecurity-focused AI model designed specifically for vetted security teams.