So Apparently AI Might Kill Us All.

Well. That escalated quickly.

☕ Afternoon Talent Intelligence — Apocalypse Edition

I spend a lot of time thinking about AI.

How it changes recruiting.

How it changes jobs.

How it changes productivity.

How companies should prepare their workforce.

Apparently, I may have been thinking too small.

🤖 Evan Hubinger, an alignment researcher at Anthropic, told CBS News that he believes there's a greater than 10% chance AI could "kill all humans" within the next decade.

Not take our jobs.

Not automate recruiting.

Not eliminate middle management.

Kill Humans.

All of us.

Within 10 years.

So there's that....

In fairness, the actual issue is much more serious — and nuanced — than the headline.

Hubinger isn't saying today's chatbot is plotting humanity's demise.

The concern is about the possibility of future superintelligent AI systems becoming capable of recursively improving themselves faster than humans can understand or control them.

And researchers working closest to these systems are openly debating whether we actually know how to keep something significantly smarter than us aligned with human interests.

That's worth paying attention to.

But I also have questions. (and I've been waiting for this article too!)

💭 Brett's Extremely Scientific AI Risk Assessment:

• AI improves productivity? 👍 Excellent.
• AI helps us understand talent markets? 👍 I'm listening.
• AI eliminates administrative work? 👍 Sign me up.
• AI becomes recursively self-improving superintelligence? 😬 Let's put a governance framework around that.
• AI kills everybody? 👎 Going to have to mark that one "Does Not Meet Expectations."

So maybe the future-of-work conversation needs one additional bullet:

• ☑️ Upskill the workforce
• ☑️ Redesign jobs
• ☑️ Establish AI governance
• ☑️ Maintain human oversight
• ☑️ Ensure continued existence of human civilization

You know. Basic change management.

And suddenly my concerns about AI writing terrible job descriptions feel considerably less urgent.