Source https://towardsdatascience.com/why-do-ai-projects-fail-9b07f32ce321

Gartner, HBR estimates up to 85% AI projects fail before or after deployment, which is double the rate for software.

AI is harder to deploy than software. AI has indeterministic outcomes. AI experiences capability uncertainty in the hands of users. Unintended consequences post-deployment results in bad press and loss of user trust in AI systems. The costs and ROI are hard to justify upfront.

Collected from sources spanning years of academic & industry research, detailed reasons below:

(1) AI systems don’t solve the right problem
(2) AI innovation gap
(3) AI systems can’t achieve good enough performance and are not useful
(4) People miss low-hanging fruit
(5) AI systems don’t generate enough value
(6) Ethics, Bias, Societal Harm