Google is great at gamifying machine learning

The latest trend on social? In the US, it’s comparing yourself to 70,000 works of art and sharing your best match. Sounds fun, right? However, there’s a lot more cool stuff going on in the background than you think.

Google has been experimenting with gamifying AI algorithms to gather data for a while (who else lost at least an entire morning to Quick Draw last year?) and it’s recently put its photo recognition learnings to use to enhance user experience. This can be most easily seen in Google Photos, which can recognise faces, show you all the photos you’ve got with smiles in, or show you all the snaps you’ve taken of your pets (which is a sure fire way to cement the fact you take too many photos of your pets…).

Image showing Google Photos search for cats
Google Photos search showing ‘cats’

There’s a race on to be the digital platform with the most advanced facial recognition tools available. Google’s main rival in the race, Facebook, wants good facial recognition tools so it can personalise your experience within the social network by suggesting photo tags and expanding its use of lenses on Instagram and Facebook Stories. The platform has previously reported that its tool is 97% accurate, just 2% short of Google’s reported 99% accuracy.

Putting that into context, their easy access to opted-in user data means both of their tools are vastly more advanced than law enforcement’s – reports claim the FBI’s facial recognition tool is only 85% accurate.

Like Facebook, Google wants to enhance user experience – we can only assume the company wants to further improve its photo search functionality, and it seems gamification as a surefire way to harness the power of the internet to do so.

Gone are the days when you’d altruistically let SETI use your computer to search for alien life forms, now you’re letting Google use your selfies to improve your search experience. And if it gives you a few minutes of hilarity to share with your friends whilst improving your internet experience, then why not?

Latest Posts

D2C has a channel problem Why platform roles and better creative are replacing the old channel plan Direct to consumer brands don’t need more social channels in the plan. What’s needed is a clearer ‘platform stack’ (sorry not being nerdy, but this is the best term I can think of!).
Read More
Snapchat for B2B. No, we’re not joking – and no, we won’t apologise for the poor joke attempt in the title. The US platform says that it is the ‘new destination for B2B marketing’. A bold statement. But is it backed up by data? Well – sort of. But also…
Read More
AI promised time back. It lied If you’ve switched AI on and somehow feel busier, you’re not imagining it. You’re now managing a tool, training it, checking it, and explaining it to everyone else. The day job still exists. That’s why we ran our “Thank fck, practical AI for marketers”…
Read More