MIT Study Warns of AI Hype vs. ROI: 95% of AI Projects Show No Profit Gains

An MIT report reveals that, despite $44 billion being invested in generative AI during the early part of 2025, 95% of corporate AI projects fail to drive revenue growth, with only 5% delivering meaningful returns. Organizations face security and integration hurdles, as well as a growing “verification tax,” since employees must double-check AI outputs that are prone to confident errors. Worker skepticism is running high, with approximately 62% viewing AI as overhyped. Some companies have reversed layoffs and rehired staff. Researchers have also identified an “AI-AI bias,” where models tend to prefer machine-generated content over human-generated work. Financial forecasts predicting trillions in economic gains fuel fears of an unsustainable bubble. Industry leaders, such as Sam Altman, acknowledge the hype but insist that long-term disruption will be significant. At the same time, proponents explore smaller language models and safety measures to salvage productivity promises and operational practices.

The Economic Times, ET Online – MIT study shatters AI hype: 95% of generative AI projects are failing, sparking tech bubble jitters – 1st September 2025