AI Leaders’ Devious Plan

Is artificial intelligence poised to destroy humanity? This is the (manufactured) debate capturing attention this week.

Controversy ignited after an Anthropic researcher quit, accusing major firms of “gambling with our lives.”

The trigger was this post on X (Twitter):

The tweet exploded in popularity, racking up 157 million views, 756,000 likes, and 155,000 retweets—making it one of this year’s most viral messages.

Initially, it seemed like a genuine grassroots stand. Good on Jacob for standing firm on his principles.

But then the intensity of the reactions escalated.

Next, Anthropic’s head of alignment (~safety) shared Jacob’s post while adding a startling probability of risk.

Source: X

“We really do earnestly believe AI could kill all humans! I personally think it is [greater than] 10% chance within the next decade.”

What?! The enthusiasm in his tone is striking. And this comes from Anthropic’s current alignment leader.

This perfectly illustrates the shape of a well-orchestrated PR campaign. The aim: terrify the public and push through favorable laws. Let’s review the proof.

Exhibits A, B and C

Just yesterday, Anthropic published a 154-page report outlining malicious actors attempting to exploit their AI systems.

The company claims bad actors tried to use Anthropic’s Claude model to:

Help Yemen’s Houthis develop hypersonic missiles
Enable China in tracking Muslim Uyghurs
Allow Russia to hack Ukrainian targets
Attempt genetic engineering of viruses

Anthropic asserts they “disrupted every operation listed.” The word “disrupted” is curious here—they refrained from saying thwarted or stopped. Keep that in mind.

The message suggests that the world’s continued stability hinges on their stringent safety protocols.

A common theme from major AI companies is that only their particular firm can be trusted with creating superintelligence. They portray all others as reckless threats to society’s survival. This narrative is nearly unanimous among these tech giants.

Then came another major development, via Reuters yesterday:

OpenAI pushes for mandatory national AI safety rules

“The prospect of AI-accelerated AI development demands more than voluntary commitments. The United States needs mandatory, capability-based national regulation that can evolve as the technology does,” OpenAI Chief Global Affairs Officer Chris Lehane said in a blog post.

…OpenAI is urging Congress to adopt capability-based national AI safety requirements, including testing standards, independent assessments, cybersecurity protections and incident-reporting rules for the most advanced AI systems.

Things are coming into focus. Some might think, Well, that sounds like a responsible approach! Who can oppose reasonable safety regulations?

Here’s the catch: these won’t be straightforward rules. Instead, they’re likely designed to stomp out smaller contenders and outlaw affordable open-source models. Privacy online could also suffer.

An FDA for AI

The majority pushing for AI regulation support creating an “FDA for AI.”

The government would test new AI developments to ensure safety. Companies would face hefty expenses to certify models and monitor users. Many startups simply couldn’t cover these costs.

Without intervention, this framework would cement a few dominant players—mainly OpenAI, Anthropic, and Google—in control.

A similar strategy was deployed by tobacco companies in the 2000s. Philip Morris backed FDA oversight, securing a major edge.

Now, every new tobacco product must clear FDA hurdles, making it tough for new challengers to emerge.

This has benefited existing companies greatly. It’s one reason I hold shares in Altria (MO), Philip Morris’ US branch. The business generates strong profits and pays generous dividends. Legal barriers limit competition.

If the same pattern takes hold in AI, it would harm the nation profoundly. Innovation would stall and overseas rivals with friendlier policies could dominate.

To Be Clear…

Some regulation for AI might be necessary eventually. But handing the reins to industry giants to write the rules is a terrible idea—history shows why.

They will close the doors behind themselves, creating a comfortable monopoly. They’ll reap massive profits and gain unprecedented influence over society and government.

AI rules must enable startups to compete fairly against the powerhouses. Open-source models must be allowed; otherwise small firms will be forced to pay monopolistic fees.

This vision is preferable. Yet, we may end up with the FDA-style approach—resulting in less rivalry, diminished innovation, inflated costs, and dangerous control.

What else lies ahead? Likely a digital ID system. This would mark the end of online privacy I mentioned earlier.

Using advanced AI might require a government-issued digital ID linked to your real identity, reducing misuse. Eventually, biometric authentication—like eye scans or fingerprints—could become mandatory to access the internet. In my view, this is inevitable soon after a major AI-powered cyberattack.

Also, OpenAI and Anthropic plan to go public soon. The timing of this entire campaign appears strategic.

Open-source AI threatens the big players, who seek to eliminate this competition. Will they succeed? Uncertain. However, our political system is vulnerable to these maneuvers. With vast sums at play, influence is potent.

More updates to come.