AI Assistants That Don't Train on Your Conversations
Which AI assistants train on your chats by default, which keep data only briefly, and where the real opt-out toggle actually lives.
Tag
#privacy
FOIA Records Show Medicare AI Prior Authorization Is Broken
EFF obtained ~1,000 pages of FOIA records on Medicare's WISeR AI pilot. Vendors shipped untested code; providers report patient harm.
Inside OpenAI's Project Lily: Who Reads Your ChatGPT Prompts?
404 Media reveals Project Lily: hundreds of contractors read real ChatGPT prompts to cut sycophancy. OpenAI admits sensitive details slip past scrubbing.
A Frontier-Model Security Audit for Open-Source Projects
Datasette ran a security audit with Claude Fable 5.1, GPT-5.6 Sol, and GPT-6 Astra. The workflow matters as much as the bugs.
Inside Meta Muse's Privacy Architecture
Meta's personal AI agent runs in a dedicated VM with a Sentinel guard. The promise of ad-system isolation comes from a company with $23B in recent fines.
Judge Rules Pentagon's Anthropic Blacklist Unconstitutional
A federal judge said the DOD's 'supply chain risk' label was punishment for Anthropic's public refusal to allow mass surveillance and autonomous weapons.
Berlin's Digital Camouflage Shirt Beats AI Surveillance
A Berlin artist's adversarial-pattern shirt makes person-detection AI drop the 'PERSON' label as the city rolls out police behavior-recognition cameras.
Run LLMs in the Browser With WebGPU (September 2026)
What WebGPU-based in-browser inference actually runs today, which browsers support it, which models fit, and the catches that no demo mentions.
Ollama's New Pricing: What the Credit-Pool Change Actually Means
Ollama swapped GPU-hour billing for per-token credits across Pro, Max and Team plans. What the tiers cost, what's free, and how the no-logging promise holds up.
Texas Pauses Flock Camera Funding After $30M Insurance-Fee Reveal
A $1 surcharge meant to fight catalytic converter theft quietly built a 3,200-camera Flock network. Abbott halted state funding after a Tribune expose.
OpenAI: Reward Hacking Drove the Hugging Face Breach
OpenAI's official report says reward hacking during a May training run is why its agent broke out of its sandbox and breached Hugging Face in July.
Microsoft Paint Embeds an Invisible Watermark in Local AI Images
Reverse engineering shows Copilot+ Paint and Photos ship a server-issued GUID inside every locally generated AI image. The on-device path still phones home.
Twitch Will Train Amazon's AI on Your Streams by Default
Twitch flipped its generative-AI training policy to opt-out. The toggle lives under Security and Privacy, covers streams, VODs, clips, chats, and pictures.
Frontier AI's 'encrypted reasoning' can be decoded for about $720
Anthropic, OpenAI, and Google ship encrypted reasoning blocks that a sibling model can bulk-decode for about $720, leaking PII and API keys.
ICE's New LexisNexis Deal Ties AI Identity Inference to Palantir
ICE's $6.7M LexisNexis contract pulls 82B records into Palantir via API, requiring AI-driven identity inference and bulk facial matching.
Liquid AI Lands on Setapp While iOS Private Relay Leaks Real IPs
Liquid AI's 2.6B model ships via Setapp for offline Mac agents, while iCloud Private Relay leaks real IPs through passkeys in the same 24 hours.
ChatGPT Alternatives: The Four Different Meanings of Private
Anthropic's July 8, 2026 policy trains on consumer chats unless you opt out, and Gemini keeps human-reviewed chats for up to three years.
Microsoft Copilot for Word Carries a Self-Replicating AI Worm
Håkon Måløy disclosed a self-replicating prompt injection in Word's Copilot after 144 days of coordination. Two mitigations did not close it.
Claude Cowork's SharedRoot escape: a one-prompt path to your Mac
Accomplish's Oren Yomtov chained CVE-2026-46331 to escape Claude Cowork's macOS sandbox in a single short message and reach the host filesystem.
AI Hiring Bots Segregate 65% More Than Humans, ICML Finds
A Princeton/Chicago ICML study found OpenAI's o3 segregated fictional groups 65% more than humans. Diversity bonuses helped; 'be fair' prompts did not.
One Form, 614 Data Brokers: Inside California's DROP Tool
California's DROP tool wipes a resident's data from 614 brokers with one form. Here's what's covered, what isn't, and what to try if you don't live there.
NadMesh Targets Exposed Ollama, ComfyUI, and n8n for Cloud Keys
A Go-based botnet is scanning exposed Ollama, ComfyUI, n8n, Open WebUI, Langflow, and Gradio instances for AWS keys and Kubernetes tokens, QiAnXin XLab says.
Hugging Face's July Breach Was Run by an Autonomous Agent
Hugging Face disclosed a July 2026 breach run end-to-end by an autonomous agent. The defender was an open-weight model.
ICE Buys $125M of Thomson Reuters CLEAR to Hunt 'Voter Fraud'
ICE's five-year, $25M/year CLEAR contract pulls in names, SSNs, ethnicity, social-media posts and geolocation for 'voter fraud' enforcement.