Key artificial intelligence news in the 20th week of 2026

Google Threat Intelligence Group described how artificial intelligence is being used to support phishing attacks, malware development, intelligence activities, vulnerability research and identity-data theft. According to the report, a case was identified for the first time in which the exploitation of a zero-day vulnerability was probably created by artificial intelligence. State-linked actors associated with China and North Korea use artificial intelligence to search for vulnerabilities, while a Russia-linked group hides the malware CANFAIL and LONGSTREAM behind code generated with artificial intelligence. The Android backdoor PROMPTSPY was also described as independently executing commands through the Gemini API. AISI reported that the ability of artificial intelligence models to solve cyber-related tasks has increased rapidly in recent periods, but reliability problems remain for more complex activities and evaluation is complicated by a lack of metrics. Microsoft’s MDASH system found 16 Windows vulnerabilities, including elevation-of-privilege and remote-code-execution vulnerabilities, which were fixed in the May security updates. South Korea adopted new laws addressing deepfakes.