Your ChatGPT conversations can now stay hidden

ChatGPT adds chat locks, but privacy has limits

ChatGPT now lets users lock individual conversations with a four-digit PIN, Face ID, or Touch ID.

The feature moves protected conversations into a separate Locked Chats section, keeping them out of regular chat history and search results. Users can still return to these conversations whenever they want.

There's a useful detail here. Locked Chats also have separate memory, meaning information shared in those conversations won't automatically appear in regular chats.

I think this is a practical addition, especially for people who use ChatGPT to discuss personal matters or share a phone with someone else. A locked conversation is useful when another person is holding your phone and scrolling through the app.

There are limits, though. Locking a chat doesn't add end-to-end encryption, change data retention, or erase information ChatGPT previously remembered.

The feature is rolling out across Free, Go, Plus, Pro, Business, Enterprise, and Edu plans.

As ChatGPT becomes more personal, will users understand the difference between hiding a conversation and protecting its underlying data?

The US DeepSeek ban now covers third-party AI tools

The US government wants DeepSeek removed from federal systems, including tools built using its AI models.

The Department of Homeland Security has ordered civilian agencies to remove DeepSeek products and third-party services using its models, citing concerns over sensitive data reaching China and potential security vulnerabilities.

The concern is serious. DHS believes government employees using these tools could expose sensitive information, even through services that don't carry the DeepSeek name.

I think the third-party restriction is the bigger development here, because agencies now have to examine the AI models running inside their software, rather than simply checking which apps are installed on government computers.

Several US agencies had already restricted DeepSeek, but this order extends the restrictions across civilian government systems.

It also comes as the FCC prepares for an October 29 vote on banning Chinese laboratories from testing electronics for the US market.

That raises a practical question: how many government tools are already using DeepSeek models without agencies knowing?

OpenAI wants a bigger role in how biotech companies work

OpenAI has acquired Orchestra Bio, a four-year-old startup building AI agents for biotech companies.

The startup helps drug developers manage operational tasks such as tracking finances and coordinating clinical trial schedules. It had reportedly raised around $20 million, and the acquisition price remains undisclosed.

OpenAI plans to integrate Orchestra’s team and technology into Rosalind, its life sciences AI model, and Rosalind Workbench, which helps researchers plan experiments and analyze results.

The acquisition makes sense. Drug research involves a lot of administrative work, and reducing that workload could make OpenAI more useful to pharmaceutical companies beyond answering scientific questions.

What stands out to me is OpenAI's decision to acquire an existing workflow product instead of building every feature internally. With Moderna and Novo Nordisk already named as Rosalind customers, OpenAI has a starting point for selling these capabilities.

The bigger question is whether biotech companies will trust one AI provider to handle both their research processes and the operational details behind them.

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.