Google's recent data is astonishing: with the help of AI tools, the team fixed as many as 1072 security vulnerabilities in just two versions, Chrome 149 and Chrome 150, released in June. By comparison, the total number of vulnerabilities fixed in 23 Chrome versions released over the past two years was only 1036 — AI has brought a qualitative change to the efficiency of vulnerability discovery and repair.
Gemini enters the scene, and vulnerability hunting enters the industrialized era
Doug Turner, Chrome Engineering Director, stated that large language models have fundamentally changed the economic landscape of cybersecurity, transforming vulnerability discovery into an automated and industrialized operation. By applying models like Gemini, Google has been able to fix vulnerabilities in advance, staying ahead of its opponents in the cybersecurity battle.
This trend is becoming an industry consensus. Microsoft announced earlier this month that it fixed a record number of 570 security vulnerabilities in its monthly updates and directly attributed the achievement to the support of AI tools. Apple, however, has not seen similar explosive growth; independent statistics show that the number of vulnerabilities fixed in 2026 is roughly the same as previous years.
When security researchers warned that AI would cause an exponential increase in vulnerability discovery, Google and Microsoft are proving with actual data that the defense side can also use AI to fight back. In this new normal where the speed of vulnerability discovery has increased tenfold or even a hundredfold, whoever can use AI to plug the gaps faster will seize the true initiative in this cyber warfare battle.
