A new AI model, Foresight, is generating buzz and debate for its potential to revolutionise healthcare in England. Designed to assist the NHS in predicting disease risks and hospitalisation rates, the model is trained on 57 million medical records from patients.
What is the purpose of this ‘Foresight’ AI?
Foresight taps into data from both hospital and family doctor records across England. Its creators envision a future where AI helps clinicians make faster, more accurate decisions, potentially saving lives through early intervention. But not everyone is convinced that the benefits outweigh the risks.
Privacy advocates and some researchers have raised red flags about the sheer scale of data involved. Even the developers of Foresight acknowledge that while safeguards are in place, there’s no guarantee that the system won’t inadvertently expose sensitive patient information.
The concern is heightened by the fact that the AI draws from such a vast pool of personal health data. We’ve seen this historically with other AI’s like ChatGPT, where people found out how to circumvent its safeguards by simply entering a certain prompt. In the context of medical records, you can imagine the risks.
The Foresight project began in 2023 with an earlier version built using OpenAI’s GPT-3, famous for powering the first version of ChatGPT. That pilot model was trained on 1.5 million anonymised patient records from two London hospitals, laying the groundwork for today’s much larger system.
Public Opinion
Public opinion on AI in healthcare appears cautiously optimistic. At a February NHS conference on AI, Nell Thornton of the Health Foundation presented findings from a large survey published in mid-2024.
The survey included over 8,000 members of the public and 1,200 NHS staff. A slight majority of the public (54%) and a more substantial share of NHS staff (76%) supported using AI for patient care. Enthusiasm was even higher for non-clinical tasks like managing appointment notes.
Still, trust remains a crucial issue, especially among older adults. Most respondents, particularly those aged 65 and over, said they wanted to be informed whenever AI is used in their care, even for something as routine as generating appointment letters.
As AI becomes more embedded in healthcare, models like Foresight promise exciting advancements- but also raise important questions about data privacy, transparency, and trust. The coming years will be pivotal in determining how these powerful tools are integrated into everyday medical practice.
We hope you’ve liked this blog and stick around to see our future releases. We cover everything from recent IT News to knowledge base articles. Stay safe!
