Published January 24, 2020
AI VS. AI HYPE
Baked into the definition of AI is a human perception of “smartness”, which is subjective and evolving. The AI Effect suggests that once AI has achieved something previously unachievable (e.g. computer beats chess champion) then it becomes old hat and is no longer considered to be AI.
Conveniently that means there is no correct answer to the question of what AI is, other than a point-in-time assessment of the views of the collective minds of the world! A philosophical rather than a scientific challenge.
And while those close to the technology might debate whether if-then-else algorithms, propensity models, expert systems and/or deep learning fall in or out of the AI bucket, the reality is that the outcomes generated might be completely indistinguishable to the end user.
Econsultancy and Adobe’s 2019 Digital Trends Report highlights “There’s a buzz around cutting-edge technology: 36% of larger organisations are now using AI, particularly to enhance data analysis – 50% more than last year”.
How should we interpret that apparently dramatic increase? It’s certainly tricky if you take the same view of the AI Effect as its formulator Tesler that “[artificial] intelligence is whatever machines haven’t done yet”.
Is it “AI” whenever a computer is faster than a human in finding a pattern in some data? It certainly sounds useful irrespective of whether it ticks the AI box or not.
Realistically one might wager that the question people are really answering in this kind of survey is more along the lines of: “Is your organisation using smart technology to enhance data analysis?”. And “smart” could mean a lot of different things depending on perception.
So, the next time you see a product or service or solution that is presented as being AI enabled or powered, try swapping in the word “smart”: AI tool becomes smart tool, powered-by-AI becomes powered-by-smarts.
And that will probably then prompt some questions around exactly what makes it smart, and indeed smarter than what exactly?