Steam is now using AI to recommend games

Steam is now using AI to recommend games

AS HUMANITY SEEMS intent on outsourcing its thought process to AI, then why not let it loose on brain downtime too? Steam creator Valve has launched a new experiment that’ll help decide what game you should buy next based on machine learning and how much you’ve played each game in your library.

For those of us with huge unplayed libraries, its main suggestion should probably be “finish these before spending more money, you spendthrift wazzock”, but that’s not hugely in Valve’s interests. Instead, Interactive Recommender will analyse “billions” of game sessions, your play history and “other salient data” to pick out the next title it thinks you’ll love.

If you don’t like what’s on the list, you can filter things out from entire genres to games made more than three years ago. You can also flip a popularity slider, letting you brag about your deep indie games knowledge to anybody who’ll listen.

This is one of three new experiments currently taking place in Steam Labs. The other two are Micro Trailers and Automatic Show.

Micro Trailers offers six-second clips of every game on Steam, allowing you to get a taste of the title before clicking through and watching a full trailer – like Vine for games.

Automatic Show is for those with longer attention spans: a half-hour programme, automatically generated every day to showcase the “latest and greatest games.” Valve describes it as “like one of those cable shopping channels without the super-absorbent chamois cloths,” which doesn’t sound hugely encouraging. Especially if you like chamois cloths.

But then all of these Labs features are experiments which may or may not become permanent fixtures, so give them a go and see if any of them help. Or you could actually tackle that Steam backlog and finish something instead…

Original Article

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.