SUDDENLY, everyone is getting worried about artificial intelligence (AI). The country’s top schools have warned of the “very real and present hazards and dangers” being presented by AI. The so-called godfather of AI, Dr Geoffrey Hinton, has resigned from Google, prophesying that the technology could up-end the job market. Another senior AI researcher warns of the possibility that a super-intelligent machine could wipe out humanity.

AI, which has progressed at a spectacular pace recently, allows supercomputers to “scrape” vast amounts of data from the worldwide web and process it with breath-taking speed to generate seemingly original content.

Its sophistication is impressive. One chatbot recently scored 96 per cent on an A-level maths paper. Apparently, half of all computer code is now AI-generated. A woman in South Shields was freed from bowel cancer after AI spotted problematic tissue unseen by the human eye.

But it has a dark downside, which has led to predictions of a dystopian future. AI turns out to be biased on race, gender, and other sensitive areas, because the titanic amounts of web-text that it trawls contains a range of undesirable factors — from hate speech to points of view that exclude marginalised people and places.

Power imbalances and inequalities, unfair treatment, and discrimination are perpetuated and even amplified. So is the worldview of affluent white males in the United States and Europe. In 2015, Google had to apologise when its AI identified a black couple as gorillas. There are fears that banks or governments using AI might reject certain groups unfairly for loans or benefits.

AI technology can be used to manipulate videos or recordings to create so-called deepfakes that show individuals doing or saying things that they never did. More insidiously, AI chatbots have been shown to make mistakes or make things up outright — a phenomenon known in the industry as “hallucinating” — which they can do with convincing authority.

This week, I asked ChatGPT: What are the theological shortcomings of AI? It responded with a sibylline remark about human dignity and the imago dei, and said that AI raised questions about the nature and uniqueness of human beings. But then it came up with some specious gobbledygook about human and divine creation, transcendence and immanence, and even ventured into eschatology (which makes a change, I suppose, from dystopia).

It was more helpful on ethics. Questioned about the ethical difference between the sending of a battalion into heavy enemy fire by an AI system or a human general, it had the simulated humility to accept that it lacked “the capacity for moral reasoning, empathy, and understanding of ethical nuances that come with experience and emotions”. Phew.

Politicians have been talking at the G7 summit in Hiroshima of the need for “guardrails” around the development of AI. But there is little concrete agreement between governments on how to police it.

Far more promising was the reaction of the pupils of Wimbledon High School, who have been experimenting with ChatGPT as a coursework aid. AI-generated essays, they say, are articulate and fluent, but they can lack originality and ideas. AI answers are unreliable, but helpful for learning how to structure arguments. “It will be a tool in our futures,” one student said. The answer is to “use it through proper critical thinking”. Nothing artificial about that intelligence.