The BBC recently had a feature on how Facebook’s algorithms were causing problems. The article suggests Mr Zuckerberg may have taken a while to realise the algorithms that determine the categories used in Facebook’s advertising have been allowing advertisers to target some unsavoury groups. It looks like he now concedes he may have just created a bit of a monster which is tricky to control.

Some thoughts:

• Is this an early example of robots taking over the world? OK a bit dramatic but I think it’s easier than we realise to accidentally create monsters if we don’t keep our wits about us.
• Do we need to adhear to special rules to stop this thing from happening? Probably. Perhaps it’s time to start using Asimov’s laws for robotics when we write our algorithms (well, certainly the first two).
• Where in the process do you insert the human component to regulate how this works? We should always build in a shim where humans get a chance to veto and correct decisions made by algorithms. This is tricky in situations like Facebook where the automation provides the scalability that allows the business to succeed.

I don’t envy you Mr Zuckerberg.