Aggregator

Pentagon's Anthropic Fight Draws Rebuke From Ex-DOD Leaders

DataBreachToday.com
3 months 1 week ago
Former Officials, Tech Groups Say Anthropic Designation Is Illegal - and Dangerous
Former U.S. defense and intelligence officials argue the Pentagon's designation of Anthropic as a supply-chain risk was politically motivated and legally flawed, warning it could erode trust in government contracting and weaken the defense AI ecosystem.

6 Lessons Security Leaders Must Learn About AI and APIs

Security Boulevard
3 months 1 week ago

Most organizations treating AI security as a model problem are defending the wrong layer. Security teams filter prompts, patch jailbreaks, and tune model behavior, which is all necessary work, while the actual attack surface sits largely unexamined underneath. That surface is the API layer: the endpoints AI systems use to retrieve data, call tools, and [...]

The post 6 Lessons Security Leaders Must Learn About AI and APIs appeared first on Wallarm.

The post 6 Lessons Security Leaders Must Learn About AI and APIs appeared first on Security Boulevard.

Tim Erlin

After Mythos: New Playbooks For a Zero-Window Era

The Hackers News
3 months 1 week ago
When patching isn’t fast enough, NDR helps contain the next era of threats. If you’ve been tracking advancements in AI, you know the exploit window, the short buffer that organizations relied on to patch and protect after a vulnerability disclosure, is closing fast. Anthropic’s new model, Claude Mythos, and its Project Glasswing, showed that finding exploitable vulnerabilities and subtle cracks
The Hacker News

Mercor 4TB 语音样本被盗

Solidot
3 months 1 week ago
Mercor 是美国一家 AI 初创公司，主要业务是为其他 AI 公司提供专家帮助训练模型和聊天机器人。它招聘专家/合同工时要求对方提供护照或驾照扫描件、自拍和录制一段语音。本月初勒索组织 Lapsus$ 披露它从 Mercor 窃取了 4TB 语音样本。语音样本加上身份证件，引发了合同工们身份被盗用的担忧。已有至少五名合同工对 Mercor 提起了诉讼，指控该公司以训练数据的名义收集语音特征，但并未明确说明这些特征是永久性的生物识别标识符。现有的语音克隆技术只需要 15 秒钟的清晰参照音频，而 Mercor 要求合同工录制的语音长度达到了 2-5 分钟，足够实现语音克隆。

Managed ad