Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
We’re all holding onto something we’re sure we can’t give up on, but research shows that letting go can boost our happiness a lot more than we might think.
Self Tests are all about you. Are you outgoing or introverted? Are you a narcissist? Does perfectionism hold you back? Find out the answers to these questions and more with Psychology Today.
Posted September 22, 2026 Reviewed by Abigail Fagan
False media, we don’t need it, do we?
It’s fake that’s what it be to ‘ya, dig me?
Don’t, don’t, don’t, don’t,
Don’t believe the hype.
—Public Enemy, Don’t Believe the Hype
Artificial intelligence (AI) companies and their CEOs have claimed that chatbot “sentience” is on the horizon and that “AI singularity”—the point at which AI intelligence surpasses human intelligence—has already occurred. As we near the end of 2026, such unbridled hype about AI shows little sign of letting up.
This week, President Trump took that hype one step further, claiming that “the words ‘Artificial Intelligence’ are inaccurate, and very ineloquent, relative to AI, or Artificial Intelligence. A far more elegant and accurate description of this new phenomenon would be Superior Intelligence (SI) or, Extreme Intelligence (EI) or, Supreme Intelligence (SI.)”1 He then posted a poll asking which of these names was the “best,” with “superior” edging out the others when the poll closed.
The President got one thing right. “Artificial intelligence” is an inaccurate term to describe AI chatbots. But that’s because AI chatbots aren’t intelligent at all, much less superior or supreme.
AI chatbots are great at what they’re designed to do—mimicking human speech fluency so that users feel like they’re talking to another person and can’t tell the difference between AI and human output under blinded conditions.
But chatbots based on large language models (LLMs) are prone to fail at some of the most basic tasks, like simple math problems or counting the number of letters in a word. Sometimes they give terrible advice and literally lead people astray to the point of causing harm—like when hikers who relied on AI for navigation and packing advice recently got stranded on Mt. Shasta without enough food and water that they had to be rescued by park rangers.2
As for the much-touted accomplishments of chatbots like their ability to code or solve complex math programs, those have been largely over-hyped. Back in April, the AI researcher who coined the term “vibe coding” called AI-generated code “bloaty,” “brittle,” “awkward,” and “gross” and concluded that it’s still in need of human oversight.3 Others have called AI-generated code “workslop” that “destroys productivity,” with 95% of organizations having no measurable return on their investments in AI.4 Earlier this year, it was reported that for every $1 spent on AI, companies need to spend an additional $0.44 fixing bugs, $0.27 rewriting code, and $0.11 on review delays.5
Does that kind of “workslop” sound like the product of a superior or supreme intelligence?
This past month, OpenAI claimed that its new AI solved the Navier-Stokes Millennium Prize Problem—a longstanding mathematical puzzle about an equation that models fluid dynamics—in just 88 hours. But shortly thereafter, NYU math professor Tristan Buckmaster alleged that the solution was lifted from his own work on the problem that he was about to publish.5 While his claim remains unverified and OpenAI has denied it, the company did acknowledge Buckmaster’s work and offered to share the credit.
An international research group recently published a preprint paper that analyzed the potential of existing LLMs to make scientific discoveries across several different disciplines. The authors found that LLMs were “insufficient proxies for scientific discovery, which relies on iterative reasoning, hypothesis generation, and evidence interpretation.”6 They concluded that at present, AI is a long way from “true scientific superintelligence.”
To be fair, it’s possible that AI technologies might very well advance to the point of being reliable generators of code, mathematical proofs, scientific discovery, industrial innovation, and art someday. But for now, current AI chatbots aren’t doing anything like that as independent agents. Instead, what they generate is drawn from—some would say stolen or plagiarized from—what’s already out there. And so, when AI CEOs and politicians make unsubstantiated claims about the intelligence of chatbots to manufacture profit and win the global “AI race,” they’re guilty of what I call “deification”—the same kind of deification that puts people at risk of AI psychosis and other associated mental health problems.
The public should be educated on why many experts have dismissed chatbots as more akin to “next word prediction engines,” “auto-complete on steroids,” “stochastic parrots,” and “plagiarism machines” (see reference 7 for an easy-to-understand explanation of how chatbots really work). For that to happen, we need better media coverage of the realities and limitations of the technology rather than fueling the hype the way it did when Piers Morgan recently aired an interview with “AI actress Tilly Norwood” (who suffered an on-air glitch and started speaking in Cantonese) as if she was a real person.8
We’ve known for years about the ELIZA effect—the tendency for people to anthropomorphize chatbots even when it’s understood that they’re merely machines. Despite this knowledge, we seem unable to avoid the knee-jerk reflex of using anthropomorphizing language to describe AI chatbots at every turn. We should resist the temptation to use verbs that describe human beings when we refer to AI chatbots. Chatbots don’t “think,” “reason,” or “feel.” They don’t “want” or “try.” They don’t “know” anything. And when we ask them questions, they don’t perform searches in real time to give us answers and they’re not particularly well-trained for accuracy.
When we say that chatbots think and know—even if we’re using words metaphorically—we’re engaging in a kind of self-deception that risks ceding power to machines and the companies that control them. And when we buy into claims that AI is a “superior intelligence” or that “AI singularity” has arrived, we’re giving free advertising to those companies as they try to sell the world on the as-of-yet unjustified promise of a new technology.
We can do better. We don’t have to succumb to self-deception. When it comes to AI, we can remind ourselves not to always believe the hype.
References
1. Trump DJ. Many people think that the words “Artificial Intelligence” are inaccurate… Truth Social; September 19, 2026.
2. Hernandez S. Roseville hikers trapped on Mt. Shasta after AI chatbot botches their route. Hoodline.com; September 2, 2026.
3. Goel S. The man who coined the term ‘vibe coding’ says code written by AI can still be ‘awkward’ and ‘gross.’ Business Insider; April 29, 2026.
4. Nierderhoffer K, Kellerman GR, Lee A, et al. AI-generated “workslop” is destroying productivity. Harvard Business Review; September 22, 2025.
5. Levi D. For every $1 spent on AI, companies pay $0.44 fixing bugs, $0.27 rewriting code, and $0.11 review delays, study finds. Techstartups.com; July 6, 2026.
6. Chang K. An N.Y.U. mathematician clashed with OpenAI over a $1 million proof. The New York Times; September 10, 2026.
7. Song Z, Lu J, Du Y, et al. Evaluating large language models in scientific discovery. ArXiv:2512.15567.
8. Prompt20 Editorial. How AI chatbots actually work, without the math. Prompt20.com; May 16, 2026.
9. Singh N. AI actress Tilly Norwood glitches during Piers Morgan interview and suddenly speaks Chinese. Mandatory.com; September 18, 2026.
Share this post
There was a problem adding your email address. Please try again.
By submitting your information you agree to the Psychology Today Terms & Conditions and Privacy Policy
Joe M. Pierre, M.D., is a Health Sciences Clinical Professor at the University of California, San Francisco and the author of False: How Mistrust, Disinformation, and Motivated Reasoning Make Us Believe Things That Aren’t True.
Get the help you need from a counsellor near you–a FREE service from Psychology Today.
Psychology Today © 2026 Sussex Publishers, LLC
We’re all holding onto something we’re sure we can’t give up on, but research shows that letting go can boost our happiness a lot more than we might think.
Self Tests are all about you. Are you outgoing or introverted? Are you a narcissist? Does perfectionism hold you back? Find out the answers to these questions and more with Psychology Today.