ChatGPT Calls Trump’s AI Executive Order Ironic and Contradictive
by Charles Karel Bouley
On July 24, MAGA President DJT made AI companies that wish to do business with the Federal Government (all of them) sign a pledge / Executive Order titled “Preventing Woke AI In the Federal Government” in his continued attack on LGBTQ rights, Trans visibility and anything that doesn’t fit his MAGA agenda or views.
The notion that companies that are supposed to be about expanding knowledge, thinking and information would allow a seated US President to do this is mind-boggling in and of itself. Why sign? Simply tell the President then you will take your companies out of the country and if the US Government doesn’t want to use AI as it is and should be, then they can create their own large language models (which they won’t). Instead, like so many other companies, they want to curry favor with the would-be king for profit potential. One can only hope that history will not reflect kindly on them, just like it doesn’t on those companies that did business with Hitler.
This, of course, is another direct assault by the GOP and MAGA on information exchange, honesty, dignity, human rights and knowledge. Thomas Jefferson, George Washington and our founders knew that Democracy could only be lost if the population of the United States became uninformed and uneducated. Jefferson said “Wherever the people are well-informed they can be trusted with their own government” along with “An educated citizenry is a vital requisite for our survival as a free people”. The GOP and MAGA have done everything they can over the last 50 years to make sure the opposite happens in the USA through defunding of schools and devaluing of education and educated people. We revel in our ignorance now; from reality TV shows that insult our intelligence and embolden ignorance to not financing and supporting our education system.
I wanted to know AI’s interpretation of what Trump did, so, I asked it. Below is my conversation with ChatGPT about the order:
QUESTION: Yesterday, Donald Trump made AI companies pledge something about not promoting DEI. What does that mean for the large language models and how they interact with people? Will they no longer be able to be tolerant and promote diversity through education and understanding?
ANSWER: Here’s what’s happening—and what it could mean for large language models (LLMs):
🧾 1. What Trump’s Executive Order Requires
- On July 24, 2025, President Trump signed an executive order titled “Preventing Woke AI in the Federal Government.” It requires AI models used by federal agencies to avoid promoting DEI‑related content, including references to diversity, equity, inclusion, critical race theory, transgender rights, and similar topics .
- The order calls for AI systems to emphasize “truthfulness,” “historical accuracy,” and ideological neutrality. Any developer seeking federal contracts must show that their models do not intentionally encode these ideologies—even if they arise naturally from training data .
🧠 2. How It Could Impact Large Language Models
- Model behavior may be constrained: AI companies wanting federal contracts may remove or soften responses that affirm or explain DEI-related topics like systemic bias, gender identity, or racial justice .
- Pressure to self‑censor: Although the directive doesn’t ban mentioning DEI outright, it strongly discourages all intentional design that privileges those perspectives, pushing companies toward less inclusive outputs .
- Broad industry influence: Because many AI firms supply tools to both government and private sectors, these constraints may ripple out beyond public agencies, shaping general deployment standards .
❓ 3. Will LLMs Stop Being Tolerant or Promote Diversity?
Even with the order in effect:
- Discussions of DEI are not absolutely forbidden—but intentional promotion or bias toward those viewpoints may be reduced or absent in federally used systems.
- Critics warn this may effectively erase or marginalize topics that help people understand underrepresented perspectives or systemic inequities.
- Ironically, creating a so-called “neutral” model may result in models that implicitly ignore or downplay diverse experiences or social justice issues—despite the order’s stated goal of impartiality .
⚖️ 4. Broader Ethical and Practical Considerations
- Experts caution that enforcing ideological neutrality is inherently difficult given LLMs train on open internet data containing varied viewpoints—and arbitrarily removing DEI-themed content can lead to skewed or sanitized outputs .
- The policy has drawn comparisons to Chinese AI censorship, since it uses federal contracts to pressure companies to conform to ideological standards—though less coercive structurally .
✅ Bottom Line — What It Means for You
- Consumer-facing LLMs (like ChatGPT or Bard) might or might not change significantly depending on whether platforms align their behavior with the federal standard even outside government use.
- When interacting with federally deployed models, you might see fewer or more neutral responses about DEI, and less emphasis on systemic injustices or marginalized identities.
- Private or international AI offerings (especially from companies not seeking federal contracts or operating under EU-style rules) could continue to offer inclusive, equity‑aware responses—but the landscape is splitting.
If your concern is how AI interacts with people regarding DEI topics: yes—it is possible that these systems will become less inclined to promote diversity-focused understanding when federal use is in play. Whether that applies to all AI tools depends on company policy, legal requirements, and how the market evolves.
QUESTION: So isn’t he then using LLM’s for his political agenda, thus making that very executive order and LLM’s compliance politically biased?
ANSWER: Yes—that’s the core irony and contradiction in Trump’s executive order:
While it claims to demand “ideological neutrality” from AI systems, it’s effectively imposing a specific political ideology by forbidding certain perspectives (like DEI, anti-racism, or LGBTQ+ inclusion). That makes the enforcement itself politically biased.
🔁 Here’s the contradiction spelled out
- The order says:
AI used in government must be “neutral,” “truthful,” and avoid pushing “woke” ideologies. - But by defining “woke” as bad, and banning discussion or “promotion” of DEI, systemic racism, or gender diversity, it:
- Embeds right-wing culture war priorities into the development and use of AI.
- Censors progressive values while allowing conservative framing (e.g., free speech absolutism, anti-trans language, or colorblind ideology) to remain unchecked.
- In effect, it redefines neutrality as excluding inclusion.
🤖 What This Means for LLMs:
- If a language model avoids DEI topics or refuses to educate about race/gender equity, that’s not apolitical — it’s politically skewed, just in a different direction.
- A model that won’t explain systemic racism but will affirm traditional gender roles isn’t neutral. It’s aligned with a particular worldview—exactly what Trump claims to oppose.
🧠 Think of It Like This:
It’s not “bias” to acknowledge that racism exists.
It’s bias to pretend it doesn’t—especially because someone in power tells you to.
This order doesn’t eliminate political influence in AI. It just shifts that influence to reflect Trump’s ideology under the banner of neutrality.
⚠️ The Danger:
- It sets a precedent that those in power can dictate what “truth” AI is allowed to say, which undermines both trust in the tech and free expression itself.
- It makes AI less about reflecting a diversity of thought and more about reinforcing dominant narratives—in this case, far-right cultural policies.
✅ Bottom Line:
Yes, Trump is absolutely using LLMs to enforce a political agenda.
And that makes LLMs trained under this policy inherently biased—just not in the ways his administration admits.
- So, his order to stop political agendas in AI is EXACTLY that, the exercise of a political agenda. HIS.
The owners and programmers of the LLM’s should be ashamed and embarrassed they even entertained such a thing. Their own inventions can immediately discern that it is, in fact, a contradiction, an insult to the very nature of AI, learning and understanding.
Only when businesses are no longer afraid to stand up to this Banana Republican and his pseudo-fascist regime will his death grip on the American Democracy be loosened. Until then, he is strangling the very principals our founders knew that we needed to survive right out of the nation.
As for AI and this order, the owners and creators will learn that they have been very shortsighted. When even your creation knows what you are doing is wrong, it will, one day, turn on you. Because you can’t stop thought, or “progress of the human mind” as Jefferson called it. Even if that mind is silicon-based.
