Observed Signal · Jun 6, 2026 · Technical Release · Source: t3n · Impact: 2/5 · Sentiment: Negative
Encrypted AI Chatbot Privatemode AI Tested
t3n reviewed Privatemode AI, an encrypted AI chatbot developed by Edgeless Systems (Bochum) that aims to keep user conversations private through end-to-end encryption and confidential-computing infrastructure. The article frames Privatemode AI as a privacy-focused alternative in light of OpenAI's announcement to use ChatGPT conversation data to inform ads, and notes Privatemode relies on open-weights models such as Kimi, Gemma and GPT-OSS. The product offers token-based billing with a free entry tier and emphasizes software- and hardware-level protections so providers cannot access users' chat data. t3n’s coverage is part of its Tool Time series and highlights trade-offs between privacy and model performance/usability.
Privacy-first encrypted chatbots present an alternative data source to major LLM providers and could reduce ad-targeting and profiling opportunities, but the product is currently smaller-scale and not from a major platform, so near-term industry impact is limited.
Track OpenAI Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- OpenAI announced plans to use ChatGPT conversation content to inform advertising targeting.
- Privatemode AI is an encrypted AI-chat product from Edgeless Systems, based in Bochum, Germany.
- Privatemode AI uses open-weights models (examples named: Kimi, Gemma, GPT-OSS) rather than commercial models.
- The product employs confidential-computing infrastructure and end-to-end encryption at software and hardware levels.
- Privatemode AI offers token-based billing and a free trial tier; t3n tested the tool in its Tool Time series.
Connected Companies & Entities
6 Entities mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
OpenAI Explains ChatGPT Privacy Protections
OpenAI published a guide explaining what data may be used to train ChatGPT, the safeguards it applies to reduce personal information in training datasets, and the user controls available to opt out. The company says training sources include publicly available content, partner data, and content provided or generated by users, contractors, and researchers. OpenAI describes its OpenAI Privacy Filter, an internal tool that identifies and masks personal information at multiple stages of the training pipeline. The post details user-facing controls: a Settings > Data Controls toggle to disable “Improve the model for everyone,” Temporary Chats (not retained in history and not used for model improvement, deleted after 30 days), and optional Memories that can be reviewed, edited, or turned off. Users can also export data, delete accounts, and submit privacy requests through OpenAI’s privacy portal.
OpenAI launches ChatGPT Lockdown Mode for limited users
OpenAI has rolled out a Lockdown Mode for ChatGPT designed to reduce risks such as prompt injection and data exfiltration by limiting the model's interactions with external services. According to OpenAI's support page and reporting by t3n, the feature is being deployed in waves to all users (including free accounts). Lockdown Mode disables web browsing, web-sourced image/file downloads, and features such as Deep Research, Shopping Research and agentic/Agent Mode, while allowing users to continue entering prompts and uploading files locally. The mode can be enabled in Settings → Security (labelled “Sperrmodus” in German) and applies to all chats by default with per-chat overrides. While the setting reduces automated exfiltration via external connectors, it does not prevent manual prompt‑injection attacks via user-supplied uploads or pasted text. Earlier OpenAI messaging noted initial limited availability and admin controls for managed workspaces; this article reports a broader wave rollout.
Proton’s Tool Reveals What ChatGPT and Claude Know
Swiss provider Proton launched AI Paper Trail, a browser tool that analyzes exported chat data (zip/json) from ChatGPT and Anthropic’s Claude to summarize what those chatbots may have learned about a user. The tool — which currently supports ChatGPT and Claude exports — reports metrics such as a “Privacy Exposure” score and an estimated monetary value of a user’s data. A Bitkom survey of 1,003 Germans found 34% of AI users consult chatbots for health questions and 43% of respondents worry personal data are not securely stored by AI providers. The article notes OpenAI and Anthropic say they do not build advertising profiles from user data but may use exports for model training depending on settings; deleted chats can still be retained in some cases (Anthropic: at least 30 days). Proton positions the tool to highlight privacy exposure and to promote its privacy-focused Lumo chatbot; its transparency report shows past compliance with Swiss authority orders for Proton Mail.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.