Healthcare is where the "most exciting" opportunities for artificial intelligence (AI) lie, an influential MP has said, but is also an area where the technology's major risks are illustrated.
Greg Clark, chairman of the Commons Science, Innovation and Technology Committee (SITC), said the wider adoption of AI in healthcare would have a "positive impact", but urged policy makers to "consider the risks to safety". He said: "If we're to gain all the advantages, we have to anticipate the risks and put in place measures to safeguard against that."
In the NHS, AI is currently being used to read X-rays — such as mammograms — helping doctors to speed up decision making and giving them more time to spend with patients, as well as helping clinicians diagnose strokes faster. Elsewhere, researchers are looking into how AI can be used to predict the damage long-term conditions such as diabetes can cause in a patient's body.
The technology can also be utilised in drug discovery, giving researchers access to vast amounts of data to speed the process up.
Mr Clark told the PA news agency: "One of the things we were struck by is how medicine is becoming increasingly personalised. Quite a lot of drugs have failed clinical trials on a broad-spectrum basis, but now you can identify more closely who they are more suitable for. They can then be deployed straight away to help save lives."
Science and Technology Committee Report into AI
His comments came as the SITC published an interim report on its inquiry into AI governance, which opened in October.
The document outlines 12 risks posed by the technology and provides a guide on how policy should be shaped to overcome them.
Mr Clark added: "If you can consider that AI can be given the ability to develop new medicine, it also has the ability to develop toxins."
Among the risks highlighted by the SITC was AI's potential to perpetuate "unacceptable" societal biases, as well as its potential to share personal information or generate material that "deliberately misrepresents someone's behaviour, opinions or character".
Debate is also ongoing about how AI would access large datasets, which are held by few organisations, or who would be liable if AI used by a third party were to cause harm.
"A lot of recommendations depend on the data on which AI is trained," Mr Clark said. "If you're conducting medical research on a particular sample or ethnic minority, then the data on which AI is trained may mean the recommendations are inaccurate." He added that no one risk included in the document was a priority and that they "all have to be addressed together".
Expansion of AI Use in the NHS
The Government has backed research on how AI could benefit clinicians with almost £150 million in funding, while NHS England chief executive Amanda Pritchard promised in June that more uses of the technology were "on the horizon".
In July, the National Institute for Health and Care Research (NIHR) released a collection of 10 studies on AI to demonstrate where it could be used in the future. The pieces of research were published between 2020 and 2023, and explored how AI could be used to detect heart disease via "smart stethoscopes" or predict the progression of diseases such as the bowel condition ulcerative colitis.
But the SITC has called for more urgency from the Government when it comes to developing policy on AI governance. "You need to drive the policy thinking at the same time as the tech development," Mr Clark said. "If the public lose confidence and are spooked by AI, then there will be a reaction standing in the way of some of the benefits."
He added that the UK had done this in the past with the Warnock Report on fertility treatment, which was published in the 1980s at a time when in vitro fertilisation (IVF) was a controversial topic.
Mr Clark said "approaching it in an early and rigorous way" through the report would mean the subject "doesn't have the contention and division it has in some other countries". He added: "You can instil public confidence in areas that are controversial."