Why AI Is Like Drawing An Owl

There’s a meme on the internet about drawing an owl. Here’s a PG-rated version:

In case you’re unfamiliar with it, here’s a very capable explanation from ChatGPT:

What has that to do with AI?

Well, it seems obvious that the latest large-language technologies like ChatGPT and Bard can be used to transform business. And that’s certainly true for tasks such as summarizing documents, which these technologies do astoundingly well.

But when it comes to using artificial intelligence in more real-world, end-to-end business processes, things start looking a lot more like the meme.

Step one is getting access to the AI technology. Step two is actually do something with it to improve your business. And just like sketching an owl, that’s a BIG step that requires a lot of skill.

Giving an organization access to AI won’t automatically lead to better business any more than giving somebody a pencil makes them a better artist.

The AI engines are increasingly the easy part, just like drawing the circles. The hard parts are turning that into something that looks more like the real thing. That requires getting the right data in the first place, actually using the algorithms as part of a scaleable, reliable, compliant, ethical, usable business process, and then ensuring that people actually use it!

What’s interesting about LLMs is that they can actually potentially help with some of what’s still hard in the process — potentially helping corral data into meaningful combinations faster and more reliably, for example, or helping business people understand how they can make best use of the new technology. But the key is that these areas will require deep business skills and knowledge about people, not so much about the details of the algorithms — so there’s still hope for us humans!