While artificial intelligence offers tremendous benefits in fields like aviation, experts warn that it could lead to “human stupidity” if we rely on it without thinking. The call for the principle of “trust but verify” has never been more urgent in the face of hallucinations and misleading content.
Article contents:
- The delicate balance between benefit and risk
- Artificial intelligence in the service of aviation safety
- Expert warning: never assume AI is correct
- Types of artificial intelligence: not everything is the same
- The danger of hallucinations and “AI slop”
- Conclusion: don’t let AI think for you
- Frequently asked questions
The delicate balance between benefit and risk
Artificial intelligence is a powerful tool transforming our daily lives and offering immense benefits to society. However, AI tools must be used with a clear understanding of potential pitfalls. Excessive reliance on these technologies without critical thinking can lead to what some describe as “human stupidity,” where users stop engaging their minds and accept machine decisions as absolute.
At the 2025 U.S. Aviation Leadership Summit hosted by Honeywell Aerospace, key leaders discussed the advantages and risks of using AI in aviation, as well as in daily use, highlighting the urgent need for caution.
Artificial intelligence in the service of aviation safety
In fields like aviation, artificial intelligence has proven its ability to save lives. Transportation Secretary Sean Duffy opened the summit, highlighting progress in improving aviation safety and efficiency. He noted that the U.S. Federal Aviation Administration (FAA) is now using AI to search the national airspace system and identify airports with potential unsafe flight operation hotspots.
Duffy said: “The human eye misses these things. If you’re reviewing data, or if multiple people are reviewing data, they don’t see the trend lines. AI does a fantastic job of helping us identify those trend lines and giving us a warning to take action.” This demonstrates how AI can be a crucial early warning system.
Expert warning: never assume AI is correct
Congressman Jay Obernolte, the only member of Congress with an advanced degree in artificial intelligence, concluded the event with a stern warning. Obernolte explained that AI encompasses “things that a machine does that look human-like to people,” changing the way we interact with automation on a human level rather than a technical level.
However, he emphasized a fundamental rule: “Never, ever, ever assume that AI is correct. Always follow the principle of ‘trust but verify.’ AI is incredibly powerful. But it is not infallible. And it never will be, because… it is trained on the breadth of all human information. And human information is fallible.”
Types of artificial intelligence: not everything is the same
It is important to understand that there are different types of artificial intelligence. Aviation strives to use deterministic AI, where a specific input always yields the same expected output, which is essential for safety. In contrast, probabilistic AI can analyze incomplete information to predict trends.
Meanwhile, generative AI, such as ChatGPT, uses patterns to create new content. However, the result may not always be as expected, producing errors or outright “hallucinations.” Obernolte explained that we should never assume generative AI output is 100% correct, and it must always be verified.
The danger of hallucinations and “AI slop”
Generative AI becomes a problem when its outputs are assumed to be correct. Numerous examples of this have emerged. Obernolte pointed to “widely publicized cases of lawyers submitting AI-generated legal briefs that [contained] entirely fabricated legal history.”
In another example, a health committee published a report containing hundreds of references to scientific studies, many of which turned out to be nonexistent or to contain incorrect information—they were simply AI hallucinations.
There is also a growing volume of so-called “AI slop” being published online. AI-generated articles and images are published as news or analysis, often filled with errors and absurdities that may seem plausible. Even reputable publications have fallen into the trap of publishing AI-assisted articles without sufficient review by editors.
Conclusion: don’t let AI think for you
Artificial intelligence breeds human stupidity if it encourages users to turn off their brains. Excessive reliance on AI without critical thinking can expose human ignorance and put us at risk. In aviation, lives depend on it.
Obernolte warned: “It is always up to us to verify what comes out the other end, because we never know if it’s going to be 100% accurate, or if it’s going to be a hallucination or something in between. The mistake we make is when we allow AI to do our thinking for us.”
Frequently asked questions
Q: How can artificial intelligence generate “human stupidity”?
A: This happens when people blindly rely on AI outputs without questioning or verifying them, leading to a decline in critical thinking skills and the acceptance of false or fabricated information (hallucinations) as facts.
Q: What is the principle of “trust but verify” in the context of artificial intelligence?
A: It means using AI as a powerful tool to assist with analysis or content creation, while always needing to verify the results and ensure their accuracy before taking any action or relying on them, especially in critical decisions.
Q: What is an “AI hallucination”?
A: It is information that looks plausible and convincing but is entirely incorrect or completely fabricated, generated by a generative AI model, such as inventing legal cases or nonexistent scientific studies.
