OpenAI takes phishing more seriously

ChatGPT now needs account security

OpenAI is treating ChatGPT accounts like places where sensitive material actually lives.

The company launched Advanced Account Security, an opt-in protection program for ChatGPT users. It also partnered with Yubico on two co-branded security keys, the YubiKey C NFC and YubiKey C Nano.

This is overdue. People are putting drafts, legal notes, research, source material, and private questions into chatbots, often with less care than they give a work email inbox.

The physical key matters because phishing is a boring problem that keeps working, and a small USB device can block a lot of that risk. OpenAI is clearly aiming this at journalists, researchers, dissidents, elected officials, and enterprise users with valuable information sitting in chat histories.

There is a catch. Lose the key, and OpenAI says it cannot help recover the account, which means conversations may be gone for good.

The key is small.

The question is whether regular users will accept that kind of responsibility for a chatbot.

Apple under-called Mac demand

Apple expected steady Mac sales, then AI buyers started clearing out Mac mini and Mac Studio units.

In Q2, Mac revenue hit $8.4 billion, beating expectations in the low $8 billion range. Sales rose 6% year over year, helped by new products like the MacBook Neo and demand for local AI work.

That matters because Apple’s AI story has mostly been judged through iPhone features and services growth. The more interesting signal is that developers and companies may already see the Mac as practical AI hardware they can put on a desk today.

Cook said supply could take several months to catch up, and the Mac mini became the top-selling desktop in China during the OpenClaw rush. That is a real demand cue, even if quarter-over-quarter Mac revenue stayed flat.

The shelves moved fast.

Now Apple has to prove this was more than a bright March spike.

The forest alerts are getting smarter

Rangers in Guatemala found a fresh 2-hectare clearing about eight days after it happened.

Now a new project in the Maya Biosphere Reserve will use bioacoustics devices trained to detect chainsaws, gunshots, engines, and other signs of environmental crime. The work is backed by the $100 million AI for Climate and Nature Grand Challenge, with winners receiving as much as $2 million.

This is the kind of AI conservation use that feels practical because it starts with a real field problem. Rangers cannot patrol 2.2 million hectares every day, and a small device in a tree can catch sounds they miss.

The hard part is not the model. It is the mud, the stolen devices, the false alerts, and the fact that some sites take days to reach.

A truck got stuck.

If this works, it will likely work best alongside camera traps, drones, satellite imagery, and ranger judgment. The real test is whether the people on the ground trust the alerts enough to act.

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.