Featured
Why 80% of AI-projects fail: they miss the bigger picture
There are no AI-problems, there are real problems where AI can be part of the solution. The reason why most AI-projects fail is because we forget that technology is there to solve our problems, not become them.
“80% of AI projects fail — twice the rate of failure of IT projects that do not involve AI.” .. “Misunderstandings and miscommunications about the intent and purpose of the project are the most common reasons for AI project failure” .. “Successful projects are laser-focused on the problem to be solved, not the technology to solve it” — Rand National Security Research Division (1)(2)
The Rand summary of why 80% of AI projects fail is simple: We are ignoring the bigger picture!
We are reducing the world to simple technology-problems ignoring everything else that makes up who we are or how we organize.
From the Rand research:
Five leading root causes of the failure of AI projects were identified:
- Stakeholders often misunderstand — or miscommunicate — what problem needs to be solved
- The organization lacks the necessary data
- The organization focuses more on using the latest and greatest technology than on solving real problems.
- Organizations might not have adequate infrastructure to manage their data and deploy completed AI models.
- The technology is applied to problems that are too difficult for AI to solve.
When AI starts taking up all the oxygen in the room the world is reduced to 1’s and 0’s, a prompting question, and efficiency gains.
But we forget that AI needs to fit into the real world, a world of humans, culture, relationships, expertise, incentives, contexts, meaning etc..
“Methodology which ignores context simply does not work” — Heung-wah Wong (3)
e.g. even if “efficiency” is our AI-goal, efficiency on its own means nothing. Do we want to drive cost reduction, risk management, improve productivity, prepare for scale, become fit for innovation etc. How or where and why are we intending to become more efficient?
It’s always a bigger question!
And once we know how, where and why we want to become more efficient we still need to figure out who contributes and how to get there:
1. Who participates, how do they collaborate?
2. Who makes decisions, what is the scope of their responsibility?
3. What do we know and how do we understand the opportunity we need to solve (and how can we improve)?
etc.
AI is crazy good at a lot of things, I share the enthusiasm. But having spent some time understanding how people and organizations work it’s unfortunate to see how opportunities are reduced to something as simple as binary problems, simple math problems (4), questions about optimization not prioritization and choice.
There is a reason why most AI-projects fail and its not necessarily because the technology isn’t there, but because we forgot to include everything else (especially ourselves) and end up asking the technology to do everything for us instead of partnering with us.
We forgot to include everything else (especially ourselves) and end up asking the technology to do everything for us instead of partnering with us.
Sources:
(1). James Ryseff, Brandon F. De Bruhl, Sydne J. Newberry, The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed, https://www.rand.org/pubs/research_reports/RRA2680-1.html
(2). Cisco AI Readiness Index 2024, https://www.cisco.com/c/m/en_us/solutions/ai/readiness-index.html
(3). Heung-wah Wong, It’s not that all cultures have business, but all businesses have culture The Routledge companion to anthropology and business
(4) Rory Sutherland — Are We Now Too Impatient to Be Intelligent? | Nudgestock 2024, https://www.youtube.com/watch?v=Bc9jFbxrkMk
