After a month of disappointing product updates, scathing op-eds, and dismal ROI, public sentiment toward AI is turning increasingly negative. Suddenly, AI boosters’ reality distortion field is looking more and more like a bubble.
Even Sam Altman thinks so. The OpenAI CEO said as much during a rare on-the-record dinner with a handful of tech journalists as he attempted to manage the fallout of ChatGPT-5, which was released at the beginning of August to widespread criticism over its less-than-PhD-level intelligence and its impersonal responses to users who’ve grown attached to the chatbot’s sycophantic tone.
“Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes,” Altman told The Verge’s Alex Heath at the dinner. “Is AI the most important thing to happen in a very long time? My opinion is also yes.”
Is AI having a very bad couple of weeks? My opinion is also yes.
Last Monday, Reuters/Ipsos concluded a survey that revealed Americans are deeply troubled about the potential effects AI could have on their lives. Seventy-seven percent of those polled said they are concerned about AI causing political chaos, and 71 percent are concerned about AI will put too many people out of work permanently.
The next day, another, California-specific study was released to similarly alarming results: 55 percent of those living in the very state where the technology is centralized are more concerned than excited about it. TechEquity, the tech accountability group that conducted the study, found that job loss, deepfakes, disinformation, and personal privacy were among the top causes for Californians’ concern.
Then The Atlantic published an essay by Charlie Warzel headlined “AI is a Mass-Delusion Event.” The piece, which was widely shared, makes the argument that AI’s main offering is “psychosis-as-a-service,” and begs the question: When a tool pumps out brain-breaking slop art, reanimates murdered children, vaporizes thousands of jobs, and breaks the hiring process, what else could we expect but widespread insanity?
It certainly hasn’t created widespread profits. This week, the media was all abuzz over an MIT study that reported that 95 percent of businesses have seen zero return on investment in AI technologies. Zero. Even Meta is freezing AI hiring while it restructures its AI division after spending exorbitant amounts to poach key developers from startups.
Then there’s AI psychosis, and chatbots having “sensual” conversations with children and deadly ones with seniors.
The data is dismal. The vibes are worse. AI is looking more like 3D TV or the Metaverse with every headline. And yet, there’s a creeping sense that none of this really matters.
“Enthusiasm from investors certainly doesn’t seem to have ebbed.” said Jon Metzler, a public policy professor in UC Berkeley’s Haas School of Business, where he teaches on competitive advantages in tech markets.
“This is a case where I’m looking at the headlines and thinking, maybe history doesn’t repeat itself, but it does rhyme a bit,” he said echoing the quote widely attributed to Mark Twain.
Metzler invoked the Gartner Hype Cycle, an oft-cited pattern plotted by analysts at the research firm Gartner that emerging technologies frequently follow: First, a boom in popularity that spurs bullish investment followed by a bust after it fails to live up to the hype.
Metzler suggests AI is in the downhill slide, careening toward a stage of the cycle called the Trough of Disillusionment.
“Often new technologies go through this, and some make it through and some don’t,” Metzler said.
To understand how corporations will proceed, Metzler suggests you have to single out enterprise AI, which is different from the many worrisome instances that are driving negative public sentiment, like a possibly nonexistent freelance journalist using ChatGPT to write articles, or a stalker using a “nudify” app. Those are individual use cases. Enterprise AI, on the other hand, refers to how AI is built into large organizations at scale.
“After a trough,” Metzler explained, “members of the industry will usually look at a peer who’s doing something valuable and say, ‘Aha!’” and mimic their techniques moving forward.
Metzler predicts there will be “a bit of a shakeout” of weaker AI organizations along with “some consolidation” of companies that are seeing stronger returns, but investment will continue, as zealously as ever. This can be seen in the studies and whitepapers made for industry insiders like the one released earlier in August from venture capital firm Bessemer Venture Partners called “State of AI 2025.” The report begins with a bolded statement:
“We should first state unequivocally that we are extremely confident that AI is driving the biggest wave of technology change we’ve ever seen.”
The hype cycle continues. It’s a new week, and the next phase of the curve — the Slope of Enlightenment — could soon be on its way.
