State of AI: Over Hyped, Disappointment, and Backlash
“People are feeling the anxiety & uncertainty, and the fragile nature of this time. The relevance around community, sense of humanity, the third place, people are longing for human connection. Take it personal. Understand your responsibility to every customer and partner. That’s how we win.”
– Howard Schultz at the Starbucks Leadership Experience
The rapidly growing disappointment with AI stems largely from a gap between what was promised and what people experienced day to day. AI was pitched as a civilization-level transformation, and the real-world version so far looks more modest. One Fortune piece frames the open question as whether AI delivers the cash flows and productivity gains in time to justify expectations or becomes a great way to rewrite your emails and put your decks together, rather than something world-altering.
Overhyping the Impact AI Will Have on Jobs
The reasoning felt airtight in its simplicity: cut costs, move faster, trim training budgets, and let AI plus a leaner workforce pick up the slack. But the actual outcome looks very different. Workloads aren’t shrinking — they’re climbing. And by gutting their junior ranks, companies have walked straight into a slow-moving demographic problem, just as older employees start retiring in unprecedented numbers.
Demographic Time Bomb
Between 2024 and 2032, an estimated 18.4 million experienced workers aged 55 to 64 with postsecondary education will leave the workforce — but only 13.8 million younger workers (ages 16 to 24) with comparable qualifications are positioned to replace them. Even as AI automates certain tasks, businesses still need people who bring judgment, context, institutional knowledge, and sector-specific expertise — qualities no algorithm can replicate.
Businesses haven’t clearly seen returns
A widely cited MIT study found that 95% of enterprises report zero measurable return on their generative-AI investments. The core question skeptics keep asking, in the words of a head of research at Goldman Sachs, is whether companies using AI actually make or save money. Analysts also argue that the idea that you can sprinkle AI onto existing processes and achieve transformation is where the hype was most inflated.
Economist Ruchir Sharma argues AI is flashing red on overinvestment, overvaluation, over-ownership, and over-leverage. There’s also concern about circular deals, with Nvidia, OpenAI, Oracle, and others appearing on multiple sides of the same transactions. One skeptic at Goldman argues that part of Frontier Labs’ record revenue exists only because they’ve been heavily subsidizing tokens, selling compute at steep discounts to build market share.
Many companies exaggerated or mislabeled their AI capabilities to attract investors, like how companies added “.com” to their names during the internet bubble. Every day, users encounter a lot of half-baked “AI features” that make products worse, souring the whole category.
Here are the most common frustrations with AI:
The loop-and-deflect pattern
Chatbots are often tuned to deflect contact volume rather than resolve issues — pointing to FAQ pages or repeating the same canned answer regardless of what the customer already tried. One consumer quoted by CNBC described “an endless loop” where the bot just points to some type of FAQs list or repeats information you’ve already tried and found lacking. Nearly one in five consumers who have used AI for customer service saw no benefit from the experience, a failure rate far higher than for AI use generally — because the problem isn’t speed, it’s whether anything actually gets solved.
No visible way to reach a human
The most common complaint pattern: chatbots or IVR trees with no clear escape hatch. Even when companies do build one in, it’s often buried. Home Depot’s 2026 voice-agent rollout is cited as a counterexample specifically because it stresses quick access to a human associate when needed — implying most rollouts don’t.
Refunds and disputes handled by bots
This is where frustration peaks — situations with emotional or financial stakes (billing errors, refunds, account issues) get routed through automation that isn’t equipped to exercise judgment or make exceptions.
Speed without resolution
Per one 2026 CX report, 75 percent of consumers had received a fast AI-driven response that still left them frustrated, about a third said AI support made things harder, and nearly 90 percent reported reduced loyalty when human support was removed — the bots hit response-time metrics while failing to actually fix anything.
The trust cost compounds
This isn’t just anecdotal irritation — it’s measurably changing behavior. 57% of customers say their trust in a business would decrease if it predominantly uses AI for customer service, 70% believe service would get worse if humans were removed in favor of AI, and 73% say they’d be more loyal to companies that keep real people for all service interactions. And frustration with AI agents has risen from 54% to 59% year over year, with more customers saying they’d hang up if connected to AI (29% to 31%).
Why it is happening despite the backlash
Internal pressure, not customer demand, is driving adoption. By February 2026, Gartner found pressure from executive leadership to implement conversational AI had risen to 91%, with nearly 80% of organizations expecting to reshape agent roles — suggesting a lot of this is cost-cutting dressed up as innovation, and it’s starting to visibly backfire: Gartner reported that half of organizations expecting to significantly cut customer service headcount because of AI would abandon those plans by 2027.
The throughline across all of it: bots optimized to deflect contact volume rather than resolving the customer’s problem, paired with no clear or fast path to a human when the bot fails. For example, a customer asked its chatbot about bereavement fares. The bot invented a discount policy that didn’t exist. Air Canada argued the chatbot was “a separate legal entity” responsible for its own words — a tribunal rejected that, ordered Air Canada to honor the fare, and the case is now cited as the precedent that companies are liable for what their chatbots say.
Broader pattern from industry research: Sinch’s 2026 research found 74% of organizations that deployed AI customer service agents have had to shut down or roll one back due to failures — and that rate was actually higher (81%) among companies with mature guardrails, suggesting the problem isn’t polish, it’s the underlying deployment logic (cutting cost vs. actually resolving issues).
Is AI Costing More than the Humans it has Replaced?
Economists remain split on whether AI will trigger a genuine “jobs apocalypse,” largely because real-world adoption is lagging well behind the pace of technological progress. A recent Brookings report found that rapid advances in AI capability are not translating automatically into broad economic gains or meaningful adoption.
Uber blew through their Annual AI Budget in just four months
Uber’s president and COO, Andrew Macdonald, echoed this in a podcast interview, saying it’s growing increasingly difficult to justify the company’s AI spending. Uber’s CTO, Praveen Neppalli Naga, drew widespread attention in April after acknowledging the company blew through its entire 2026 budget for Claude Code in just four months.
Nvidia admits AI is costing more than the employees it was designed to replace
Nvidia’s VP of applied deep learning, Bryan Catanzaro, offered a similar warning to Axios — arguing that AI isn’t cutting labor costs at all, and may actually run more expensive than the employees it’s meant to replace. He noted that for his team, compute costs now dwarf what they spend on staff.
Microsoft questions the sustainability of AI Adoption
Microsoft has voiced similar concerns. The company has reportedly started pulling back employee licenses for Anthropic’s Claude, citing the high cost of continued use — a move that raises broader questions about how sustainable AI adoption really is at scale.
OpenAI’s Sam Altman says he was wrong about the Jobs Apocalypse
Sam Altman, CEO of OpenAI, has made a series of sweeping predictions about how AI would reshape the job market. He’s claimed AI would likely take over most jobs people currently perform, that some professions would disappear entirely, and that displaced workers would eventually find new kinds of work to fill the gap. More recently, though, Altman seems to have shifted his position — saying he’s happy to have been mistaken about how disruptive AI’s impact on employment would actually be.
“I don’t think we’re going to have the kind of jobs apocalypse that some of the companies in our space advocate or talk about,” said Altman. He went on to explain that the “human part” of employment could not be replaced by AI and that people care about interacting with one another at work.
The common thread is that companies gave the bot authority to make commitments (refunds, discounts, policy statements) or public-facing autonomy without a human checkpoint, and had no fast, reliable way to catch or contain it when it went wrong.
Is your investment in AI improving Your Customer Experience, or Driving Your Customers Away?
Technology should enhance the customer experience, not replace the human connection that builds trust and loyalty. At The DiJulius Group, we help organizations create world-class customer experiences that balance innovation with meaningful human interactions.
Schedule a complimentary call with The DiJulius Group today and discover how to build a customer experience strategy that puts people first.
