Repost: Like everyone else in the room at the excellent and informative KIMRA Conference on AI in the knowledge sector, I’m in little doubt that unless you’re running full tilt at AI the bus will leave without you.
What I picked up through the day:
- The whole issue of what “value” means for clients as a professional services company is in play
- How do knowledge businesses capture that elusive, high value “tacit knowledge”
- AI is already transformative in areas where there is strong source information but otherwise .. caveat emptor.
- Analysts need to be able to do what AI *can’t* do – critical think, explore anomalies, absences, move beyond obvious inferences
- On that basis I think arts-based degrees that promote critical thinking and checking sources may well come back in vogue
- We all know we’ve got to run full speed at this, try things, break stuff and accept the disruption for what it is.
On a related note, having watched an AI tool confidently and repeatedly serve up a fabricated set of evidence for a theory about consumer behaviour yesterday, I’m minded to come back to Kathryn Schulz’s 2010 book, Being Wrong: Adventures in the Margin of Error.
In it, she describes a familiar situation where you’re asked a question and as the follow up questions continue, you freestyle answers as if you were an actual expert in the topic. Schulz affectionately calls this information that you’ve read in a fictional magazine called “Modern Jackass”.
That’s one of the big challenges about AI. It’s incredibly powerful, and will get more powerful the more we feed it. But it doesn’t “think” per se and just as the internet supercharged the availability of information, it too will make it harder to determine what is really true from what could be true. In some senses, it is Schulz’s Modern Jackass.
