U.S. AGENCY: THIS YEAR AI WILL FIND TWICE AS MANY INTERNET SECURITY LOOPHOLES AS LAST YEAR

On 28 July, the number of software safety gaps identified in 2026 among hot-tech products was expected to rise to about twice the number discovered in 2025, a surge that was due mainly to the increasing capacity of artificial intelligence systems. The United States National Gap Database shows that 45,207 leaks were recorded from January to Monday, a figure close to the total for the whole of 2025. Last year, the number of loopholes recorded in the database reached an all-time high. OrcL.N indicated that its monthly software upgrades in July had repaired 1,449 security gaps, which had reached an all-time high, compared to 309 in the same period last year. Microsoft (MSFT.O) revealed 642 security loopholes in July, also at an all-time high, almost five times the number for the same period last year. Google (GOOG.O) discovered and repaired 433 such loopholes in a recent Chrome browser update, compared to 11 of the same updates a year ago. Gabriel Shapiro, a leading artificial intelligence research scientist for Cybersecurity, said: “We must face up to the fact that these tools are enhancing people's ability to detect software loopholes.” Google Chrome Engineering Director, Doug Turner, stated that the loopholes had been discovered “on an unprecedented scale and speed”, thanks to advances in artificial intelligence models and corresponding investments。