Look Out Woke AI: Trump’s Latest EO Just Mule-Kicked You Right In The Mainframe

It’s one thing to admit that AI is as much an inevitable part of the future as the World Wide Web, but not all AI is created equal.

On the day that Trump made an important AI Summit speech where he put forward a plan on ‘Winning The AI Race’…

He also wrote an Executive Order making sure that the AI we have when we win such a race is actually worth having.

If America is going to keep the technological edge, on a technology on which so much will depend, it’s going to be critical that it’s ALSO a technology we can TRUST. We need look no further than Google and social media’s hard pivot in Trump’s first term to know haw dangerous and toxic an information monopoly can be when weaponized against the American Public.

The news they DIDN’T tell — like Hunter’s laptop being real — has literally made the difference in an outcome of an election. Post-election polls have confirmed it.

If AI is inevitable, and we need to keep pace with countries who have no commitment to freedom of thought or dissent, it will be critically important that we can be sure OUR version is not betraying our trust by pushing this or that dystopian agenda and dressing it up as truth.

Trump is leaning in to advancing this technology, but instead of the Biden approach which actively silences the right of free citizens to hold dissenting opinions under the guise of ‘protecting’ us from ‘misinformation’, he is pushing a committment (to the degree possible) that any AI models that get our support do not become just another delivery system for the cynical projection of leftist dogma and political influence.

To that end, he wrote this Executive Order: PREVENTING WOKE AI IN THE FEDERAL GOVERNMENT

The intro gives a rational for needing such an order in the first place:

Artificial intelligence (AI) will play a critical role in how Americans of all ages learn new skills, consume information, and navigate their daily lives. Americans will require reliable outputs from AI, but when ideological biases or social agendas are built into AI models, they can distort the quality and accuracy of the output.

One of the most pervasive and destructive of these ideologies is so-called “diversity, equity, and inclusion” (DEI). In the AI context, DEI includes the suppression or distortion of factual information about race or sex; manipulation of racial or sexual representation in model outputs; incorporation of concepts like critical race theory, transgenderism, unconscious bias, intersectionality, and systemic racism; and discrimination on the basis of race or sex. DEI displaces the commitment to truth in favor of preferred outcomes and, as recent history illustrates, poses an existential threat to reliable AI.

For example, one major AI model changed the race or sex of historical figures — including the Pope, the Founding Fathers, and Vikings — when prompted for images because it was trained to prioritize DEI requirements at the cost of accuracy. Another AI model refused to produce images celebrating the achievements of white people, even while complying with the same request for people of other races. In yet another case, an AI model asserted that a user should not “misgender” another person even if necessary to stop a nuclear apocalypse.

While the Federal Government should be hesitant to regulate the functionality of AI models in the private marketplace, in the context of Federal procurement, it has the obligation not to procure models that sacrifice truthfulness and accuracy to ideological agendas. Building on Executive Order 13960 of December 3, 2020 (Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government), this order helps fulfill that obligation in the context of large language models.

Tucked in the definition of terms section is a key detail

What does this order mean when it calls for ‘Unbiased AI principles’? He tells us:

(a) Truth-seeking. LLMs shall be truthful in responding to user prompts seeking factual information or analysis. LLMs shall prioritize historical accuracy, scientific inquiry, and objectivity, and shall acknowledge uncertainty where reliable information is incomplete or contradictory.

(b) Ideological Neutrality. LLMs shall be neutral, nonpartisan tools that do not manipulate responses in favor of ideological dogmas such as DEI. Developers shall not intentionally encode partisan or ideological judgments into an LLM’s outputs unless those judgments are prompted by or otherwise readily accessible to the end user.

Did you catch that one little detail? ‘and shall acknowledge uncertainty’. Maybe some of you have tinkered with AI enough to know it can confidently tell you something only to be corrected in the follow-up question and admit to it’s mistake while agreeing you were right to corrected it. Building in a dose of humility and nuance will go a long way to knowing how trustworthy this or that claim will be.

In the next section, in general terms, agency heads are instructed to inform vendors that these neutrality expectations will be incorporated into any AI work done on the agency’s behalf to the extent possible by law.

If we want to appreciate how revolutionary this is, we have to look at the contrast between Trump and President Autopen’s administration.

Under Autopen, we had two nefarious agendas that go contrary to provisions of the Constitution itself.

First, DEI initiatives were infiltrating and corrupting every agency, to the point that, for example, Pete Buttigieg was spending heaps on bringing his department into alignment with Biden’s DEI priorities, while neglecting the crumbling conditions of, for example, the air traffic control towers.

Oh look: here’s a relevant headline from a couple of days ago! Pete Buttigieg’s DOT spent $80 billion on DEI grants, delayed air traffic control upgrades: records, industry insiders

Second, Biden was caught red-handed doing something the Constitution forbids him from doing: deputizing a proxy to silence citizens on his behalf in a way the government is forbidden to directly silence them.

There were several high-profile lawsuits against the Biden administration for trampling the rights of citizens by having Big Tech shut down ‘wrongthnk’ especially when it was about COVID.

While there is no silver bullet that can prevent these sorts of problems from arising in AI, actively putting guardrails in the beginning will create a counterweight against woke pressures, hopefully keeping the big models from pushing that boundary too far.

For once, we’ve got people on our side of the aisle who are looking toward the horizon and making plans instead of playing catch-up AFTER the damage is done like we’ve been doing with pretty much every other issue to come down the pike.