Sources

I try not to say anything on this site I can’t back up. When I make a claim about the industry, like what a company did, what a court ruled, or what privacy research found, the source is listed here so you can read it yourself and decide whether I got it right.

The references are in APA (7th edition) format, grouped by topic and then listed alphabetically. Industry news moves fast, so every entry carries its date, and I recheck this page before any campaign goes out. The AI Privacy Index reports carry their own citations inside each report.

Last reviewed: November 16, 2026.


What the sources cover

Companies that restricted ChatGPT in 2023, and what happened next

In 2023 several large organizations, including Samsung and a number of big banks, restricted staff use of ChatGPT over data-leak worries. Many of them later replaced the ban with AI tools they run and govern themselves. Samsung lifted its ban on outside AI tools in June 2026.

See: Broersma (2024); Lee (2026); McGlauflin (2023); Paoli (2023).

Whether your chats with a public AI tool are protected in court

In February 2026 a federal judge in New York ruled that a defendant’s chats with a public AI tool were protected by neither attorney-client privilege nor the work-product doctrine. Courts don’t all agree. The same month, a federal court in Michigan protected a litigant’s ChatGPT work as work product, and a Colorado court followed in March. The question is still open, and the details of each case matter.

See: Crowell & Moring LLP (n.d.); Husch Blackwell LLP (2026); Kelly et al. (2026); Stanciuc (2026); Wilson Elser (2026).

Whether people read your AI chats, and what opting out does

Reporting in September 2026 described contractors reading real ChatGPT conversations to improve the model. It also found that switching off the training setting stops new chats from being used, but doesn’t pull back chats that were already eligible.

See: Arntz (2026); Cox (2026); Oh (2026).

Health data and AI: which products come with a BAA

Under HIPAA, a vendor handling protected health information needs a Business Associate Agreement. OpenAI offers one for its API and for certain sales-managed ChatGPT plans. It does not offer one for ChatGPT Business.

See: OpenAI (n.d.).

How AI providers compare on privacy

Independent researchers rank the major AI platforms on how much of your data they collect, share and use for training. Worth knowing: Incogni sells a data-removal service, so it has a commercial interest in privacy findings.

See: Incogni (2026).

The memory systems I build on

Every brain I install is open source. A Basic Transfer runs on Honcho, and the upgrade tiers can use Open Brain, OpenViking, Mem0, Letta or Graphiti. The licenses and descriptions on the How It Works page come from each project’s own repository.

See: Jones (n.d.); Letta (n.d.); Mem0 (n.d.); Plastic Labs (n.d.); Volcengine (n.d.); Zep Software (n.d.).

What “blind” processing is, legally

Handling your archive under a random ID, with the key that links it back to you kept separately, is what European data-protection law calls pseudonymisation. The data can still be linked back with that separate key, so it still counts as personal data and gets handled that way. That’s also why I call it “blind” rather than “anonymous.”

See: European Parliament & Council of the European Union (2016).


References

Arntz, P. (2026, September 15). How to opt out of AI chatbot training. Malwarebytes. https://www.malwarebytes.com/blog/how-to/2026/09/how-to-opt-out-of-ai-chatbot-training

Broersma, M. (2024, July 29). JPMorgan rolls out AI chatbot to staff. Silicon UK. https://www.silicon.co.uk/cloud/ai/jpmorgan-chatbot-ai-573263

Cox, J. (2026, September 14). Inside “Project Lily”: The humans reading your ChatGPT chats. 404 Media. https://www.404media.co/inside-project-lily-the-humans-reading-your-chatgpt-chats/

Crowell & Moring LLP. (n.d.). Crowell tracker of court rulings on legal privilege and artificial intelligence tools. Retrieved November 16, 2026, from https://www.crowell.com/en/insights/client-alerts/crowell-tracker-of-court-rulings-on-legal-privilege-and-artificial-intelligence-tools

European Parliament & Council of the European Union. (2016). Regulation (EU) 2016/679 (General Data Protection Regulation), Art. 4(5). Official Journal of the European Union, L 119. https://eur-lex.europa.eu/eli/reg/2016/679/oj

Husch Blackwell LLP. (2026, February 23). Heppner v. Claude: The first privilege waiver by AI ruling—What lawyers and clients must know. https://huschblackwell.com/newsandinsights/heppner-v-claude-the-first-privilege-waiver-by-ai-rulingwhat-lawyers-and-clients-must-know

Incogni. (2026, August 17). Gen AI and LLM data privacy ranking 2026. https://blog.incogni.com/gen-ai-llm-privacy-ranking-2026/

Jones, N. B. (n.d.). OB1 (Open Brain) [Computer software]. GitHub. https://github.com/NateBJones-Projects/OB1

Kelly, J. M., Pencook, N. J., Koonts, J., & Sherman, S. N. (2026, June 8). Everything’s bigger in Texas, including work product protection for AI chats. Nelson Mullins. https://www.nelsonmullins.com/insights/blogs/corporate-governance-insights/all/everything-s-bigger-in-texas-including-work-product-protection-for-ai-chats

Lee, J. (2026, June 12). Samsung reverses years-long ban on external gen AI use. CIO. https://www.cio.com/article/4184660/

Letta. (n.d.). Letta [Computer software]. GitHub. https://github.com/letta-ai/letta

McGlauflin, P. (2023, May 19). Apple, Goldman Sachs, and Samsung among growing list of companies banning employees from using ChatGPT at work. Fortune. https://fortune.com/2023/05/19/chatgpt-banned-workplace-apple-goldman-risk-privacy/

Mem0. (n.d.). Mem0 [Computer software]. GitHub. https://github.com/mem0ai/mem0

Oh, D. (2026, September 17). OpenAI hired contractors to review ChatGPT chats but Improve opt-out does not remove earlier eligible chats. International Business Times UK. https://www.ibtimes.co.uk/openai-contractors-reviewed-chatgpt-conversations-project-lily-1820430

OpenAI. (n.d.). Getting a Business Associate Agreement for the OpenAI API. OpenAI Help Center. Retrieved November 16, 2026, from https://help.openai.com/en/articles/8660679

Paoli, C. (2023, May 3). Samsung bans employees’ use of ChatGPT. Redmondmag.com. https://redmondmag.com/Articles/2023/05/03/Samsung-Bans-Employees-Use-of-ChatGPT.aspx

Plastic Labs. (n.d.). Honcho [Computer software]. GitHub. https://github.com/plastic-labs/honcho

Stanciuc, A.-M. (2026, April 15). A US judge ruled that a fraud defendant’s AI chats with Claude are not privileged. The Next Web. https://thenextweb.com/news/us-court-ai-chats-not-privileged

Volcengine. (n.d.). OpenViking [Computer software]. GitHub. https://github.com/volcengine/OpenViking

Wilson Elser. (2026, March 5). Divergent: Two federal courts chart seemingly opposite paths on work product protection for AI-generated materials. https://www.wilsonelser.com/publications/divergent-two-federal-courts-chart-seemingly-opposite-paths-on-work-product-protection-for-ai-generated-materials

Zep Software. (n.d.). Graphiti [Computer software]. GitHub. https://github.com/getzep/graphiti


Something wrong or out of date?

Tell me. I’d rather fix it than have you find it: info@ai-brain-transplant.com.