SECURITYWEEK’S weekly roundup highlights how AI is shifting threat actor tradecraft, with Google Threat Intelligence Group noting in Q4 2025 that adversaries are increasingly integrating AI across attack stages, including model extraction attempts and broader automation via API access to models like Gemini, though Google disrupted many of these efforts. The collection also covers a DoD employee, Samuel D.
Marcus, indicted for acting as a money mule in a multimillion‑dollar scam, alongside California’s $2.75 million fine on Disney for CCPA violations and the EPA reporting vulnerabilities at 277 community water systems. In addition, Reuters reported that the Trump administration delayed some China tech restrictions to avoid provoking Beijing ahead of an April summit.
Taken together, the items underscore ongoing AI‑assisted threat activity, enforcement actions on consumer privacy, critical infrastructure vulnerabilities, and policy moves affecting cross‑border tech risk. The piece also spotlights other notable stories, including a supply‑chain flaw affecting access to 200 airports and related cautions for security teams.