In August 2025, MIT published a viral report about the Generative Artificial Intelligence Pilots fail to deliver measurable business value. The headline spread fast—sparking skepticism about AI’s real potential.

But here’s the problem: While the report highlights real challenges, its methodology was flawed and overly negative. Critics, including the Marketing AI Institute, argue that AI adoption is rising steadily, and it’s delivering measurable ROI across industries.
The MIT Report’s Claims
95% Failure Rate: Most AI pilots fail to deliver ROI.
Focus on Generative AI: Findings were tied to generative use cases, but applied broadly in the media.
Reasons for Failure: Poor data management, lack of integration, and employees not knowing how to use tools.
The Criticism
Flawed Methodology: Small sample sizes, weak statistical rigor.
Misleading Narrative: Oversimplifies AI adoption, ignoring thousands of successful cases.
Skewed Perception: Reinforces fear instead of highlighting opportunity.
Why AI Pilots Can Fail (Melt Studio Perspective)
From our perspective at