Warning: A timely reminder that not all AI is created equal

While we are right to be aware of the concerns and challenges that AI could represent in the future – stretching from copyright infringement, to creation of deep fake content, to false reporting of tense situations, threats to democracy, and even to potential loss of employment – not all AI is created equal.

 The vast majority of conversation on this line is around “generative AI”, that is, artificial intelligence used to generate and create content based on the manipulation – wholesome or egregious – of existing content and data sets.

 But not all AI is created equal: the very capability to take and manipulate data, then model, predict and enact outcomes, can be used in a vast range of applications for good and for the benefit of users.

We’re beginning to see some of that already. For example, “narrow” AI can be applied to broadcast encoding that adapts to user demand in video streaming to reduce bandwidth and power consumption. Or, dynamic allocation of network resources to improve mobile phone coverage, and again save energy. Or, modelling of climate change to provide more precise focus on specific sustainability goals. Or, faster and safer drug discovery to cure terrible illnesses. This is the good stuff that can positively affect our personal experiences and wider world.

I’m not at Bletchley, sadly, but my hunch is we’ll see enormous focus on the bad that generative AI represents. There’ll be a lot of scaremongering and pushback.

Instead I’m strongly advocating that we don’t throw the baby out with the bathwater, and really hope our technology leaders can recognise the other forms of AI can take and how it can be used positively.

Hashtag #AIforGood as you might say.