The Midas Project

Great to see OpenAI adopting a new whistleblower policy which, while flawed in important ways, nonetheless puts the company ahead of its peers on this issue.

The industry still has a long way to go, but this is a step in the right direction.

Last August, alongside @consumerfed.bsky.social and others, we warned that xAI hadn't done enough to prevent non-consensual sexual deepfakes.

Now the EU has announced its investigation into the company.

Good to see regulators taking this seriously.

Right-wing pundits suddenly hate an AI bill. Are they getting paid to kill it? - Model Republic Suspicious similarities in posts from over a dozen conservative influencers including Laura Loomer, Brad Parscale, and Ryan Fournier hint at a coordinated-yet-slapdash effort to stop the AI OVERWATCH ...

Read the full story at our new publication, Model Republic: www.modelrepublic.org/articles/rig...

We reached out to Nvidia, Influenceable, and every reachable account mentioned in our report.

None responded.

There's no smoking gun, but each of these accounts independently gaining an interest in the minutiae of semiconductor policy on the same day seems unlikely.

Multiple posts suggested the bill would benefit China, with one saying Congress could "greenlight sales" of chips to China or help our adversaries.

The bill does the opposite. It creates a mechanism for Congress to block chip exports to adversary nations.

Various posts point fingers at democrats, with one calling the bill "sponsored by a long list of degenerate Democrats."

In reality, the bill was introduced by Republican Brian Mast. Every single cosponsor is a Republican. There are no Democrats on the bill.

Undisclosed paid influence campaigns are ethically questionable but increasingly common on social media.

A deeper problem in this particular case is that the posts are full of claims that don't survive contact with the actual bill.

On January 15, the same day these posts started appearing, White House AI Czar David Sacks quote-tweeted Wall Street Mav's post with a single word: "Correct."

The administration is listening to this campaign. That's the point.

If this were paid, who could be behind it? Who benefits from killing this bill?

One possibility is Nvidia, which has been lobbying hard to sell chips to China. CEO Jensen Huang met with Trump in December and publicly opposed the legislation.

Approximately half of these accounts have documented or apparent ties to Influenceable, a PR firm that pays conservative influencers for undisclosed posts.

We can't prove Influenceable ran this campaign, but the overlap is suggestive.

Other themes from the posts include handing control to Congress, stripping Trump of his presidential authority, hamstringing/tying his hands, and blaming the bill on Democrats.

Two posts even contained the same typo, writing “AL” instead of “AI” (It's a hard mistake to make when writing, but an easy mistake to miss when copy-pasting from a shared document.)

Two more used the exact phrase "Democrats and their Deep State partners.”

The posts weren't just similar in opinion.

They shared the same phrases, the same metaphors, and the same false claims.

8 accounts used "win/lose the AI race."
7 used "strip Trump of power."
5 used "hands control to Congress."
3 named "Hakeem Jeffries" specifically.

The AI OVERWATCH Act is a Republican bill that would let Congress review AI chip exports to adversaries like China. It's backed by Microsoft and right-leaning think tanks.

But starting January 15, influencers called it pro-China sabotage and a Democrat plot, all in unison.

Something strange happened on conservative Twitter on Thursday.

A dozen right-wing influencers suddenly became passionate about semiconductor export policy, posting nearly identical (and often false) attacks over a 27-hour period on a bill most people have never heard of.

🧵

You can learn more about OpenAI's proposed restructuring in The OpenAI Files.

www.openaifiles.org

At the same time that OpenAI makes splashy claims about supporting nonprofits, it's harassing nonprofit critics (including us) with subpoenas, asking for our documents and communications about the company.

But the media falls for it still.

OpenAI did not just create "one of the richest charities in the world."

It WAS one of the richest charities in the world, and now it still is -- but with its ownership share substantially diminished.

But now that it has succeed, OpenAI and its investors have gotten dollar signs in their eyes.

They're persuing a "recapitalization" that will substantially weaken the nonprofit's control and entitlements. The latest version of that is a measly 20% stake.

How is this possible?

OpenAI once had profit caps limiting investors to a maximum of 100x their initial investment, with all future cash flows returning to humanity.

By 2021, this dropped to single digits. The nonprofit was owed nearly dollar OpenAI made if it succeeded.

OpenAI's comms team is weirdly effective at generating headlines that make it seem like they've done an incredible thing (given $100b to their nonprofit!) while actually undercutting their past commitments (diminishing the nonprofit's entitlements significantly!)

OpenAI once said its nonprofit would be entitled to "the vast majority" and "all but a fraction" of the wealth it generates.

Now, in their new restructuring, they are saying it will be entitled to only 20%. (~$100b out of a $500b valuation).

From "Nearly all" to "one fifth" 🙄

NEW: @consumerfed.bsky.social along with 14 other advocacy groups is urging investigation into Elon Musk's xAI for their tool facilitating creation and distribution of Non Consensual Intimate Imagery.

Coverage @theverge.com: www.theverge.com/x-ai/759554/...

Letter consumerfed.org/wp-content/u...

We look forward to hearing OpenAI’s response.

✉️ To read the letter, and to add your name, visit
www.openai-transparency.org

CEO Sam Altman once asked for this sort of accountability — saying that OpenAI never wants to make decisions to benefit shareholders, but only wants to be accountable to humanity as a whole.

This letter is offering such accountability.

What is it that the letter’s signatories, spanning creatives, professors, nonprofits, and whistleblowers, have in common?

We are all part of humanity, and thus beneficiaries of OpenAI’s mission.

Everyone has a stake in this restructuring and deserves to know the details.

7. Finally, will the public ever get to see the details of our current entitlements and safeguards as they are legally instantiated in OpenAI documents, or will they continue to be quietly changed without our knowledge (as happened to the profit caps in 2023)?

6. OpenAI’s charter also includes a promise that, since a race to AGI could pose severe risks, if another company is close to developing the technology, they will stop racing and start assisting to reduce the market pressures and encourage safety. Is this still true?

5. OpenAI once boasted about its commitment to ensure AGI technology, when developed, is controlled by the nonprofit instead of investors. However, recent reporting suggests they are considering scrapping this promise, or replacing it with “ASI,” a much higher benchmark.

4. OpenAI’s leadership has said its profit caps are critical because the wealth they will generate is “for sure not okay” for one group of investors to have, and that if they succeed, “all but a fraction” will be returned to the world.

Is that still true?