top of page

AI in Education - What can we learn from the business world?

I spoke this morning to a group of business professionals, explaining some of the hard won lessons that other businesses have learned about how to be successful with AI. It occured to me that many of these learnings may apply to education. They are not really about businesses, they are about the role of people in transformation. Here are 5 hard won lessons from business, which I think apply to education

(1) If you do not know what problem you are trying to solve, AI cannot help you.

AI is exciting. There is a temptation to believe that if you add AI, it will magically make things better. Alternatively, there may be an equally unjustified fear of AI. The question is - do you have a theory of what benefit you are expecting the AI to deliver? Can you measure it? If AI improved the situation, would you know and how would you know?

(2) Iteration is key.

Businesses that are successful with AI almost always got there by iteration. The first version may or may not work. The key is to - as said in (1) - know what success is supposed to look like, measure, and if you did not get there the first time, iterate quickly and fix whatever you can think of that you can do better.

(3) Simple quickly is better than complicated later.

One of my favorite business AI failure stories is about a company that decided that their first AI would use the most sophisticated AI types they knew about. After 6 months of development, they discovered that a simpler AI could work equally well, and be a lot cheaper and easy to understand. The embarrasment was the least of their problems. Try something simple, as quickly as possible, and see what you learn about what else you need. As in (1) and (2) measure and iterate!

(4) AI is a team effort. Think about your team

Businesses that isolated AI into siloed teams often found that the AI that they created and "threw over the wall" failed in the hands of the practitioners that it was built for. This occurs because AI behavior is complex and the practitioners were not involved in its creation so they did not know how to manage its nuances. Similarly, businesses have found that getting domain expertise input from the beginning can enable creation of simpler, more robust, and more effective AIs.

(5) Upskill everyone

The AI profession has a high turnover rate. One reason is that companies hire AI professionals without supporting the rest of the organization with AI knowledge. This makes it difficult for everyone - the AI professional cannot be effective, and the rest of the team does not know enough to work with them. A mix of new input and upskilling makes an effective team (in 4).

How do these learnings apply to education? It appears that they all do. They suggest tempering the excitement and fear with solid, measurable goals, iterating quickly with measurement, and bringing everyone along for the ride.

29 views0 comments


bottom of page