What is everyone talking about this week: Can AI pass aesthetic judgement?
Someone asked ChatGPT and Gemini what the most beautiful city in Britain is. Their answer seems fair. Where does that leave us?
Ask six different AI systems for the most beautiful city in Britain and the answer is most likely Edinburgh. Chatronix.ai, a tool that purports to ‘put all your AIs in one place’, compiled a list last week based on the different answers given to the question by ChatGPT, Gemini, and four others. Edinburgh came out on top, followed by Bath, York, Cambridge and Oxford. This sounds broadly correct — although, personally, I’d bump Oxford up to bronze. (Ed: the writer went to the University of Oxford so it's surprising that he didn't bump it up to silver while he was at it.)
The methodology, however, raises two important questions. By treating AI’s answers as reliable, we recognise the value of consensus over expert opinion. ChatGPT and Gemini scan not only newspaper articles, but online forums in a bid to give a more accurate answer.
Are they right to?
The second question contends not with the results, but with the impetus for such an experiment. Increasingly, tourists are using AI to plan their holidays. I myself have done it to varying degrees of success; as ever, don’t trust everything you read. This is bad news for mid-tier travel agencies, whose clientele is more likely to cut costs on bespoke itineraries.
Is this how things end for them?
