The tech world is always looking for the next big thing: the next iPhone, the next Twitter, the next Google. Technology marches on, after all. And share prices are based on future value. The current next big thing is AI – artificial intelligence (even though it's not really artificial intelligence).
If the whole notion of AI seems a little too much for you, you’re not alone. On the Nine O'Clock Show, Brendan Courtney spoke to cognitive scientist Dr Abeba Birhane, who lectures in Trinity College, Dublin and has been appointed to the UN’s AI Advisory Board.
At Brendan’s request, Abeba gave us "AI for Dummies".
"AI is everywhere. It’s ubiquitous. You know, all over the names, all over the media, even within the research space. So, the basic idea of Artificial Intelligence – at least in its conception back in the 1950s and the 1960s – was to recreate intelligence. It could be problem-solving, it could be development, you know, it could be various cognitive faculties. The idea was to recreate those faculties in machines."
Over the decades, though, that original goal changed and nowadays, Abeba says, we call almost any automation, "AI".
We can think of the ways intelligence can be recreated in machines in sub-categories: the three examples Abeba gives are computer vision, natural language processing and robotics.
In computer vision, the idea is to map and understand the visual world:
"And the way you do that with machines is you collect thousands and millions and billions of image data, sometimes video data and you develop a formula, an equation and you feed all that data into the machine, and it will be able to kind of learn what the visual world looks like."
More machines learning than artificial intelligence, then. A classic example of computer vision that Abeba talks about is facial recognition technology. With your formula and your billions of photos of faces, you tell the machine billions of times, 'This is what a face looks like. This is also what a face looks like.’
"And by the end of the training, when you show the machine an image that it hasn’t seen before, then it should be able to tell you, you know, face or not a face."
Facial recognition is, of course, contentious from all sorts of angles, including privacy and racial profiling. And it’s something that has also hung over the whole rush to stick an AI label on any piece of tech, from your car to your toothbrush.
"So, AI has come to denote surveillance. You know, all those gadgets, the smart fridges, the nice, convenient vacuum cleaner, even the smartwatch on our wrist, it all comes as convenient and it does make life easier, but the downside of it is also that we’re constantly being tracked and monitored and surveilled as well."
Is this surveillance that much of a problem, Brendan wondered. Abeba is convinced that it is:
"We are creating a society where there is no escape, where there is no space just to be because you are constantly watched and monitored. Even the vacuum cleaner itself, even the sanctuary of the home is no longer just for your own safety, for you to just be.
"It has become, you know, these machines have infiltrated into our daily lives so much that even our home is constantly watched and monitored. So that should be a problem."
What about the fears people have expressed that AI will ultimately end up destroying humanity in a sci-fi movie-style nuclear Armageddon featuring fearsome robots that all look like a certain Austrian action star?
Turns out that’s pretty far down the list of things we need to worry about:
"A lot of the AI narrative is filled with unrealistically overhyped understanding and kind of sci-fi-like abstract things that we should worry about. So, there is this understanding that AI is this, you know, sci-fi kind of Terminator type of stuff that is going to kill all humans. Or you have this narrative of long-term impacts of AI where people think that eventually we won’t be able to control AI and that should be a worry. But all that is really unwarranted. All that is more of a theory, more of, you know, sci-fi, rather than reality."
That’s a relief.
But what about our phones? They’re full of machine learning-driven software. Are they listening to us all the time? Yes, yes, they are.
There’s lots more to get to grips with in the full chat between Brendan and Abeba, which you can listen to by clicking above.