Danish electical plug looks like a face

Large Language Models/transformers are phenomenal, and excitement around these technologies is appropriate, but good grief, how much of the discussion jumps the shark?

Why so credulous when it comes to AI? We seem to welcome all and any spurious claim.

Cognitive Bias

In part we’re led astray by our cognitive biases, the shortcuts we use to understand the world. The most relevant of these for our enthusiasm for AI is our love of anthropomorphization.

We see faces in inanimate objects, Eliza was heralded by many as an amazing psychotherapist, early Roombas were given names, and even the ancients saw emotions in natural phenomena. We all massively over-generalize from small slices of behavior or shapes that we associate with living things.

So what chance do we have at taking a cold-eyed, rational view of the realities of something that works with language as well as todays AI/large language models, let alone when there’s human-shaped robots out there.

Is it any wonder that so much of the discussion around these technologies trends to the breathless and fantastical.

Even the term carries baggage

Worse, the very term ‘Artificial Intelligence’ takes us further down exactly the same path of over expection. I much prefer the Japanese equivalent term which I’m told translates to ‘Artificial Wisdom’, which is much closer to ideas of ‘canned knowledge’ rather than sentience and science fiction world level threats.

A call for clarity and pragmatism

We need clarity of language around the realities of today’s AI. The pragmatic view, if you like. What does it actually do? Where strong and where weak? Where does it bring economic and social wins, where is it fragile, where should it be regulated, where not? What are the grounded, near term risks to our informational world?

A deeply pragmatic approach is critical when faced with the realities of emerging tech, what’s real, what’s stable, what’s defensible, what’s likely to churn, and implications for the market. This is core to what we do at Robotics Hub, and should be step one for anyone investing in advanced technologies/deep tech, especially in fast evolving fields like AI and Robotics.

__

One step towards unpacking hype is to understand the incentives behind various claims. What’s driving each commentator? Who will win from a particular narrative? And who will lose?