What would someone have to say to convince you that you didn’t need to look out the windscreen while driving? After all, you are provided with a steering wheel and a brake to control how reality plays out, so there seems to be a purpose in looking out.
Companies have spent years developing tools for handling data, and people have spent years becoming expert in data analytics. Of course they want to sell their product and their labour. Software tools for handling data are not good at handling maybes and what ifs – it requires an entirely different approach and it requires intelligence.
But these data management tools use AI!
They use a form of AI that is completely shorn of intelligence – they use training data, so every possibility must have already happened before (to be in the training data), and the future market response (savage, sluggish, trending, runaway) is not factored in, because the particular form of AI doesn’t support that. It can report trends in the data, but extrapolation is a different thing altogether. An obvious problem – have the ups and downs of inflation been put in the training data, was the war in Ukraine in the training data a year ago, or COVID before that? Shocks can make a nonsense of training data, with its trainers chasing their tails. Externalities loom large for many companies – “The answer is in the data” is like whistling Dixie – it makes you feel better, but it doesn’t do any good.
A dynamic market, with competitors trying to get in front of you, or run you off the road, is not the place to rely on data based on outdated training data.
Data management tools only have value if you recognise their strict limitations, and make sure that intelligence, whether human or machine, is factored into the mix. Of course, we are flogging an AGI (Artificial General Intelligence) tool, so we would say that.
Could someone persuade you not to look out the windscreen while driving?