The AI boom is not all hype – here’s how the UK can capitalise on it
2023 is being hailed as a ‘break-out year for artificial intelligence’, and with the launch of Open-AI’s Chat GPT, Google’s Bard, DeepMind’s Sparrow, Anthropic AI’s ‘Claude’ and, now, Chat GPT-4, it might well be true.
From a technological perspective, the launch of Chat GPT-4, one of the largest and most powerful language processing AI models to date, is a significant step forward. Its impact is reminiscent of the sudden performance of neural networks on the image net a few years ago, and then a few years later, General Adversarial Networks (GANs).
Models like ChatGPT are no doubt pushing boundaries, but they are yet to cross them. This was acknowledged by OpenAI co-founder Sam Altman who said that, while it gives the illusion of greatness, ‘it's a mistake to be relying on it for anything important right now’. In other words, don’t get too excited.
AI’s potential is huge, but at the moment, it still requires correction. And as we have seen with Google and Microsoft’s new models, these chatbots have not been error-free. This is not to minimise the accomplishment (which is huge) of reaching the stage where the AIs are as sophisticated as a teenage human, merely to remind ourselves that they are no more able to be left unsupervised than our own offspring. Yann LeCun, the famous AI researcher, puts it in harsher terms: we're not even at dog AI or cat AI, let alone human AI. I.e., even the most sophisticated AI models have less common sense that your pet.
While headline-grabbing chatbots have done a lot to bring AI into the mainstream consciousness, it is the sensible, practical applications of generative AI that are perhaps the most meaningful today. And you don’t need to look any further than the UK to see some of the most interesting applications.
Just recently, UK-based Kheiron Medical Technologies announced the use of its AI – ‘Mia’ – to help spot breast cancer markers during mammograms in a first-of-its-kind trial in Leeds. In another sector entirely, OpenAI-backed ‘Harvey’ is using generative AI to streamline legal work. In fact, a company that I helped to found back in 2017 – Luminance – has developed its own large language model that is automating the work of lawyers in over 500 businesses worldwide.
Having next-generation AI that automates the menial, repetitive tasks that plague so many professions is no bad thing. It just requires humans to learn to work and think differently, to let go of the 90% of tasks that an AI could do better and focus on the 10% that an AI will never be able to replicate. Strategic thinking and creativity to name a couple.
And while US companies seem to be dominating the headlines, there is no reason why the UK can’t be part of the story. We are home to three of the top 10 research universities in the world and world-class tech talent. With momentum around AI at fever pitch and the British government placing technology at the core of their growth ambitions, now is our moment to lean into this.
Dr Mike Lynch OBE FREng FRS