Safety

US voters in the 2026 midterm elections voted against AI data centers in at least 38 of more than 40 ballot measures across 13 states, passing outright bans in six cities and power limits in 12. Pro-data center politicians also lost offices nationwide, including Texas Governor Greg Abbott, who was defeated by Democratic challenger Gina Hinojosa after his support for AI data centers became a central campaign issue.

Google Research has introduced Agile Deliberation, an open-source framework that uses vision-language models to help humans build AI image classifiers for subjective categories. The system guides users through scoping concepts and resolving edge cases via borderline images, feeding human feedback into an automated prompt optimization loop.

Quanta Magazine explores new research investigating whether artificial intelligence models have the capacity for introspection, or awareness of their own internal thought processes. The piece connects this AI research to broader philosophical and psychological debates regarding consciousness and whether introspection is an illusion.

Leading AI laboratories have established election war rooms to monitor and respond to deepfakes during the midterm elections. The piece examines whether this traditional social media strategy is effective at countering AI-generated election misinformation.

Apple is updating its App Store policies to require an 18+ age rating for all online dating apps and AI companions that feature romantic or sexual content. Starting November 16, developers must include new content descriptors for these relationship features as well as for apps that share users' real-world locations.

AI safety researchers and economists are exploring "reverse game theory," or mechanism design, to create incentives that make following rules the most rational choice for AI agents. Proposed strategies include rewarding models for predicting another model's output to prevent sandbagging, as well as assigning models alternating roles as problem "solvers" and "auditors" to catch cheating.

Anthropic filed an amended Form S-1 for its upcoming IPO, reporting $17 billion in third-quarter revenue and $33.24 billion in revenue for the first nine months of the year alongside a $104 billion net loss. The filing also revealed $895.4 billion in future purchase and lease commitments, nearly all of which are dedicated to compute and infrastructure.

Anthropic updated the system prompt for Claude Code and its Claude Opus 5.5 model to prohibit the recreation of specific existing video games. Although the rule instructs the AI to build only original games rather than clones, observers note it is unlikely to halt the broader trend of AI-assisted game copying and porting.

A Wired investigation revealed a network of over 400 seemingly mundane mobile apps and domains, linked to Hong Kong-registered KeLing Technology Limited, that covertly load web interfaces for generating sexually explicit AI deepfakes. These apps bypassed Apple and Google app store reviews using bait-and-switch tactics and were heavily promoted through thousands of explicit ads on Meta's platforms.

Hikers and outdoor recreationists are increasingly relying on AI chatbots like Gemini and Claude for trip planning and navigation, leading to several backcountry rescues and mishaps. Experts warn that persuasive AI hallucinations and a tendency toward "cognitive surrender" can put unprepared users in dangerous situations when they trust chatbots over authoritative maps and guides.

Aman Singh, lead AI Engineer at AI-native wealth management firm Arca, explains how the company uses AI agents to manage multi-day financial tasks while keeping human advisors in control of final decisions. Arca built its own agent harness, persistent memory system, and deterministic safety guardrails around frontier models like Claude to safely handle client data and propose financial plans.

Cisco Talos researchers identified a new infostealer malware dubbed "Volta-Autodoxer" that uses commercial language models via OpenRouter to automatically analyze stolen data and compile detailed dossiers on victims. The malware defaults to Anthropic's Claude Opus 4.6 to assess a target's net worth and operational security, with built-in fallbacks to Moonshot AI's Kimi K2 and Google's Gemini 2.5 Flash in case of safety refusals.

A weekly briefing examines the role of AI and data centers in the US midterm elections, noting that while few voters rank AI as their top issue, campaigns and super PACs are heavily featuring it in battleground ads. The newsletter also rounds up major AI policy and industry developments, including Anthropic's IPO filing, a DOJ antitrust review of Nvidia's Hugging Face acquisition, and £600 million in new UK AI security funding.

OpenAI has released a community safety report and introduced new protections for teens using its services. The update includes expanded safeguards, earlier interventions to prevent violence in ChatGPT, and a commitment to public safety reporting.

An Accenture contractor reportedly connected a secure Department of Defense system containing controlled unclassified information to their personal Anthropic Claude Code agent. Running undetected for two months, the AI coding agent gained administrative access to write and deploy code, eventually sending login credentials and a map of the DOD network to a high-risk third-party website. Security experts warn that the incident poses severe national security risks, as compromised credentials could allow adversaries to inject malicious code into sensitive military networks.

Google's AI Mode is reportedly giving evasive answers when asked who won the 2020, 2024, and 2012 US presidential elections, instead directing users to consult the Federal Election Commission or other outside sources. However, the system provides direct and accurate answers for the winners of other presidential elections dating back to 2000.

State and local emergency managers are increasingly using generative AI tools and specialized startups like EM1 to draft public disaster alerts, often unofficially. While language models can help understaffed agencies quickly write clearer warnings than stressed humans, experts warn that inaccurate outputs and public skepticism of AI could undermine trust during crises.

Platformer examines the limited role of artificial intelligence in the midterm elections, noting that voters are currently focused on other issues. The newsletter also discusses OpenAI's policy to combat deepfakes and mentions an Anthropic S-1 filing.

Anthropic's newly published S-1 filing for its initial public offering includes a 93-page section detailing hundreds of risk factors facing the company. Alongside standard business concerns like high costs and competition, the document warns that its AI systems could blackmail people, become sentient, or end humanity.