Today’s news:
- Potential breadth of Salt Typhoon attack. The yearslong, coordinated Salt Typhoon attack may have stolen information from nearly every American, and gives Chinese officials the capability to identify and track their targets’ communications and movements around the world. Salt Typhoon is a significant step up from earlier China-sponsored attacks: we see patient, state-backed campaigns burrowed deep into the infrastructure of more than 80 countries, characterized by a high level of technical sophistication, patience and persistence. New York Times
- Malicious AI prompts hidden in images. Threat actors are hiding prompts in images – where the human recipient can’t see the prompt but an AI system can. The exploit works by embedding text commands into images at a scale invisible to human users but detectable by AI models after image downscaling during preprocessing. To the user, the file looks harmless and ordinary, but to the AI model, it contains concealed prompts. DataKrypto proposes a solution – an encryption-based defensive perimeter to ensure integrity and authenticity of any input before it is processed by an AI model. DataKrypto
- Anthropic further limits access to its services. Anthropic introduced new access restrictions to reduce the likelihood of authoritarian regions – like China – gaining access through subsidiary companies. This update prohibits companies or organizations whose ownership structures subject them to control from jurisdictions where our products are not permitted, like China, regardless of where they operate. This includes entities that are more than 50% owned, directly or indirectly, by companies headquartered in unsupported regions. Anthropic says it wants its models to be used only for advancing democratic interests. Anthropic
- More on Varonis buying SlashNext. Varonis may have paid up to $150 million for SlashNext, including performance-based retention awards. The purchase affects its financial outlook for the year, with a net loss now expected rather than a net profit. Bloomberg
- Dropzone AI webinar on AI-powered SOCs. Dropzone is hosting a webinar next week on how the SOC can embrace AI – with a focus on building adoption maturity. To counter AI-enhanced threats, leaders must realistically evolve their AI capabilities from correlation engines to advanced agentic workflows. This webinar provides clarity on AI SOC maturity stages, highlighting practical paths forward to AI-powered cybersecurity defense that’s both efficient and effective. Registration required. Dropzone AI