Self-driving cars don’t drink and medical AIs are never overtired. Given our obvious flaws, what can humans still do best?
By Paula Boddington – Artificial intelligence (AI) might have the potential to change how we approach tasks, and what we value. If we are using AI to do our thinking for us, employing AI might atrophy our thinking skills.
The AI we have at the moment is narrow AI – it can perform only selected, specific tasks. And even when an AI can perform as well as, or better than, humans at certain tasks, it does not necessarily achieve these results in the same way that humans do. One thing that AI is very good at is sifting through masses of data at great speed.
Using machine learning, an AI that’s been trained with thousands of images can develop the capacity to recognize a photograph of a cat (an important achievement, given the predominance of pictures of cats on the internet). But humans do this very differently. A small child can often recognize a cat after just one example.
Because AI might ‘think’ differently to how humans think, and because of the general tendency to get swept up in its allure, its use could well change how we approach tasks and make decisions. The seductive allure that tends to surround AI in fact represents one of its dangers. Those working in the field despair that almost every article about AI hypes its powers, and even those about banal uses of AI are illustrated with killer robots.
It’s important to remember that AI can take many forms, and be applied in many different ways, so none of this is to argue that using AI will be ‘good’ or ‘bad’. In some cases, AI might nudge us to improve our approach. But in others, it could reduce or atrophy our approach to important issues. It might even skew how we think about values.
We can get used to technology very swiftly. Change-blindness and fast adaptation to technology can mean we’re not fully aware of such cultural and value shifts. more>