Professor John Naughton of Wolfson College, Cambridge, who is the technology columnist of The Observer newspaper, examined the impact of artificial intelligence on the media landscape in the first of what will be an annual series of lectures under the auspices of the CJA-UK.
The softly-spoken Irishman began by redefining the terms of his brief: the one thing AI was not, he said, was artificial intelligence. It was the result of massive amounts of computing power being deployed to absorb vast quantities of written material, enabling the program to do something “very basic” in any exchange with a user: guess the next word from the context.
What had caused a “moral panic”, said Naughton, was the arrival of ChatGPT, which allowed anyone to use AI. As with the internet, which began in the 1980s but did not take off until the first browser, Mosaic, appeared in 1993, AI, or machine learning, had been around for some time before ChatGPT “hit the world like an electric shock”.
Despite the fact that machine learning was intelligent only in a narrow sense – it was neither intuitive nor empathetic, and could not only make things up but provide imaginary references for its statements – it would have a dramatic effect on what the columnist preferred to call the media “ecosystem”. He likened it to a desert dominated by “one giant species”, broadcast television.
The internet, social media and smartphones had transformed that ecosystem into a tropical rainforest, “and AI is a new beast in this jungle”. He predicted that it would drastically amplify everything that was already wrong with social media, but added that in the world of technology, it was common to overestimate the effects of innovation in the short term while underestimating them in the longer term.
Misinformation and deepfakes
While AI could significantly enhance human capabilities in the media, as in many other fields, it would also “turbocharge” misinformation and deepfakes. More jobs, and categories of jobs, extending ever further into white-collar sectors, would be under threat.
Naughton warned: “Machine learning has the potential to cause a massive increase in existing levels of inequality, where productivity increases but the gains are not shared. This is a greater danger than some super-intelligence wiping us all out.” We needed an intelligent response to AI, he concluded, rather than “just cringing in fear”.
The lecture took place at the House of Lords, hosted by Lord Black of Brentwood, deputy chair of the Telegraph Media Group, and moderated by Rory Cellan-Jones, former BBC technology correspondent. Prof Naughton was joined in discussion by Dr Richard Danbury, senior lecturer in journalism at City, University of London.
