Dangerous Assumptions In The AI Debate

We keep asking the wrong question: Will AI take our jobs?

The better question is…
What are we doing before it does? (you know… if the promise of AI is actually true).
Last week, Dario Amodei, CEO of Anthropic, put a number on it:

AI could wipe out half of all entry-level white-collar jobs in the next five years. 

Not evolve them… not augment them…
Eliminate them.

Cue the usual pushback…

“This is just hype.”
“This is just code… and it’s not replacing us.”
“Humans are still in control.”
“Tech companies always hype so their sales and valuations go up.”
“You’re being foolish if you believe that AI could ever replace humans.”
“Why would we ever build something that could destabilize the economy and replace us?” 

All fair… until it isn’t.

Because this time, the hype is the product… and the product isn’t coming… it’s already working. 
The automation of knowledge work is no longer a thought experiment.
And that – not the jobs – is what we’re underestimating most.

CEOs aren’t waiting for AGI. 

They’re quietly testing and/or working with AI now. 
Behind closed doors, they’re asking:
“Why would I pay a human to do this when AI can do it faster, better, and without needing a vacation?”
At places like Shopify and Fiverr, executives are already requiring managers to justify hiring a human over using AI.
At the same time, unemployment among new college grads is climbing – especially in fields like finance and computer science. 

Unemployment for college grads is now higher than the general rate.

This might not be connected (right now), but what happens if those lines do cross?
And this isn’t just about automating “low-value” tasks.
AI is no longer helping you write… it’s writing for you.
AI is no longer helping you code… check out vibe coding and watch it ship code.
Some call this “augmentation.”

But if the AI is doing anything over 60% of the job… what are you doing?

So what’s the plan?

A token tax on model usage to fund retraining?A universal basic income?A hiring freeze until we figure this out?

We’re not having those conversations… are we?

Not in boardrooms.
Not in politics.
And definitely not at scale beyond spaces like this (with those who are already interested in the future of work).

We’re betting that this is just another wave of automation and it will follow the same curve as previous ones… 

That jobs will “shift” (with new kinds of work, titles and businesses that will be created) not “vanish.”
But even optimists (like me) are admitting:
The speed and scale of this disruption is unlike anything we’ve seen.

And here’s the deeper risk:

If you’re training the AI with your own work… and that work becomes the benchmark… what happens when the system no longer needs you?
Or worse… when the new analytics is watching everything that you do at work with artificial intelligence and benchmarking you against everybody else… while at the same time, the better you get at using it, the more you are helping it to actually replace yourself. 
It sounds like an episode of Black Mirror, but stay with the thought… as you work with AI and get better at using it, are you actually training yourself out of work?

Does that sound so far-fetched? 

What if we’re not just competing against the machine… but onboarding our replacement?
This isn’t about fearing AI.
It’s about being honest about what’s already happening.
I’m also getting tired of the trope that “AI won’t take your job, but someone using AI will.”
What if that’s outdated too?
We shouldn’t assume AI won’t replace us.

We should prepare for what happens if it does.

This is what Elias Makos and I discussed on CJAD 800 AM.

Mitch Joel · Dangerous Assumptions In The AI Debate – The Elias Makos Show – CJAD 800

Before you go… ThinkersOne is a new way for organizations to buy bite-sized and personalized thought leadership video content (live and recorded) from the best Thinkers in the world. If you’re looking to add excitement and big smarts to your meetings, corporate events, company off-sites, “lunch & learns” and beyond, check it out.