What if the skeptics are right about AI? | DW News

Summary

The report examines the possibility that the AI boom is built on overstated technological promises and an unsustainable financial model, presenting the argument of AI critic Ed Zitron alongside caveats from the reporter.

Zitron argues that today’s large language models are less capable than their marketing suggests. Because they generate responses probabilistically, he says, they can produce useful code but also make mistakes or “hallucinate”—risks that grow when they are given complex tasks. He contrasts these limitations with claims that AI could eventually transform medicine or achieve much more.

His central concern is the industry’s economics. Citing financial reports and company disclosures, Zitron says major AI firms are spending at an unprecedented rate without proving that their businesses can become profitable. He also questions the strength of their customer demand, arguing that a significant share of revenue comes from other unprofitable AI companies. The report notes that these claims are difficult to verify fully, while acknowledging that leading companies have raised large sums and are investing heavily in computing infrastructure.

Zitron also criticizes the use of annualized run rate (ARR) figures. These estimates project a short period of revenue across a year, which can make growth appear more substantial than the company’s actual annual earnings. The report clarifies that run rate is commonly used as a measure of momentum and is not inherently fraudulent, but says critics and some investors believe it can mislead when presented without context.

If investors conclude that AI firms cannot deliver the growth or profits expected of them, Zitron predicts a severe market downturn, widespread failure among AI startups, and a sharp fall in demand for chips and data centers. He argues that AI infrastructure may be less useful for other purposes than the fiber-optic networks left behind after the dot-com crash. He also predicts that a collapse could reshape venture capital and challenge the assumption that major tech companies can continue growing indefinitely. The report stresses that these are forecasts, not established outcomes, and that other analysts expect some companies to fail while stronger firms survive.

Zitron says investors should be wary of market movements driven by AI-related hype, though the report explicitly notes that his personal views are not mainstream financial advice. He also blames media coverage for repeating striking growth claims without sufficiently explaining the underlying figures, and calls for journalists to scrutinize companies more closely rather than amplify their promises.

Presenters and contributors

• Hannah Hummel — presenter/reporter
• Ed Zitron — AI critic, author, journalist, and podcaster
• Producers: Joel Doy, Eric Nelson, Ricardo Cialva, and Sebastian Secla