The evolving role of generative artificial intelligence in healthcare was the focus of the annual Healthy Longevity Global Innovators Summit, with experts outlining its potential to help people live longer, healthier lives. The opening speech of the conference was Peter Leevice president of research and incubations at Microsoft, who had a personal connection to the topic of AI and medicine.
Lee said ChatGPT-4 helped his family manage his elderly father’s health remotely, saying the technology excelled at explaining complex medical issues. He remembers him and his sisters trying to juggle the information given to them by doctors and caregivers with little time to talk to them. The stress of the situation deteriorated the relationship between the siblings.
“Part of it was the fuel, and the fire ended up being due to our lack of understanding of these lab tests and reports,” Lee said, adding that it was empowering to give his father’s medical information and ask ChatGPT the three most important questions to ask when talking to the doctor.
“GPT’s ability to give us advice just lowered the temperature and really kept the family harmony,” Lee said.
In terms of longevity and AI, Lee expressed optimism that there are many areas where GPT-4 can help lead to new scientific discoveries, drugs, therapies and diagnostics.
“What I’m always drawn to is the empowerment of people themselves, as we live longer, and the people who care for us as we age,” Lee said. “For that, I really have a lot of hope.”
Even though the concept of artificial intelligence dates back almost a century, generative AI has progressed rapidly in a short period of time. OpenAI ChatGPT launched to the public in November last year and is already the standard for AI chatbot development. Generative AI refers to AI tools that respond to user prompts and can create text, images, and videos based on those prompts.
While AI tools like ChatGPT are colloquially called chatbots, Lee highlighted how advanced GPT-4 is compared to its predecessors GPT-3 and GPT-3.5. However, despite GPT-4’s impressive capabilities, Lee highlighted its limitations and emphasized that while human qualities can be attributed to it, it is still just a machine without consciousness.
“GPT-4 is not really human, of course,” he said. “If you think of a computer as a machine capable of perfect memory recall and perfect calculations, GPT-4 is not a computer either: it’s a new type of tool that we’re all trying to understand how to use.”
Lee also pointed out that AI has proven to excel at identifying flaws in medical studies, but introduces those same flaws when asked to write its own reports.
“GPT-4 was consistently able to outperform human reviewers in identifying bias, harmful and non-inclusive language in submitted manuscripts, in a way that human reviewers could not,” he observed. “Even if GPT-4 were asked to write the same article himself, he could be guilty of the same biases.”
Lee praised the conference organizer for taking an active role in guiding the development and use of AI in healthcare.
“I’m excited that the National Academy of Medicine, in particular, is taking the lead in this area, to understand enough that this can be a net positive benefit for us,” he said.
Major companies, including Microsoft, Google, Amazon and Meta, have invested heavily in artificial intelligence over the past year. Microsoft has formed a close relationship with OpenAI, investing $10 billion in the company and integrating ChatGPT into its Bing search engine and software suite. Office tools, including a new digital assistant, called Copilot, intended to replace its previous digital assistant, Cortana. Microsoft’s investment in OpenAI also brought AI developer Dall-E 3’s image generator to Microsoft Bing.
The collaboration between Microsoft and OpenAI has not been without controversy. In July, OpenAI disabled access to its “Browsing with Bing» for its subscription-based ChatGPT Plus service after discovering that users were using it to bypass paywalls and privacy settings on websites. In September, OpenAI reinstated Browse with Bing, the service will no longer be limited to data until September 2021.
As generative AI becomes more ingrained in daily life, these tools have been used for less esoteric purposes, such as processing large documents that otherwise would have taken human hours and allowing users to ask questions about the documents. In July, a British startup, Twins healthhas unveiled an AI platform designed to analyze MRI scans and detect diseases in their early stages.