Skip to main content

The next AI leap: the surprise

 

The current AI efforts are to offer us more of what we want. But there comes a point when we're too bored with what we want. So the next AI leap should be to surprise us.


We all know the current recommendation systems. 

  • Do you like that? Then try this. 
  • Based on your previous searches...
  • People like you do that. 
All the current recommendation systems aim to give us more of what we want. In order to do that, these systems need to know what we like or what people like us like. 

For years these recommendation systems only cared about similarity. Either similarity between content or similarity between people. This strategy worked and continues to bring results. However, quite soon similarity might not be enough to engage us. 

These days there is no shortage of content. People create more information than they're able to consume. One of the biggest problems these days for people is boredom. Yes, in the ever bigger ocean of content, how often do we feel like we're sick of it all? 

Recommendation systems are partly to blame for the boredom we feel (or is it burnout?). They're doing their job very well, supplying us with more and more stuff we are likely to want, but stripping us of chances to get surprised. 

Surprise is like a mega booster for your attention. Suddenly you're fully immersed, you're attending to a surprising event and nothing else matters. Recommendation systems need to learn how to surprise us otherwise we might get increasingly numb and unresponsive. 

How to design for surprise 

Random

Just show me an option that should not be there and see how I will react.

Random plus collaborative filtering

Compare me to another similar person who was successfully surprised and show me the same option.

How to measure surprise 

The current measurements are not precise proxies for the real feeling. It's all clicks, views, time spent. But how would you otherwise measure surprise?

I don't know and maybe we can't, maybe we just need to make our proxy metrics better. Maybe if engagement spikes - it could be considered as a "surprise". Or maybe we can ask people directly, but that is unreliable as any other interpretation. 

One thing is certain, the fight for user attention is fierce already and will get even bloodier. While creators come up with more and more content, their customers are getting more bored. Technology needs to find a way to surprise people to win their attention back. 

Popular posts from this blog

Product management and operations tools - Jira Product Discovery review

  JPD is a new player in the market of product management software. Jira (and the whole Atlassian suite) has been one of the most popular tool stacks for teams to deliver software products. Now they're adding a missing piece - product discovery.

Product Vision: an elevator pitch for your product

On this blog, I write a lot about making data-driven decisions . But what if you just starting to think about your product? You have a vague idea and nothing more. No point to go for prototyping or even talking to customers as you don't know yet who to talk to and what to talk about. In such situation - start from creating a product vision.

2 simple but powerful filters for your problem and product ideas

Nowadays lots of people and companies want to innovate. They want to generate new ideas and turn them into profitable products. But how would you separate good ideas from not so good ones? How would you make sure you invest only in good ideas?