Chatbot nurses capable of providing care for patients and passing medical exams could be the answer to alleviate NHS staff shortages, according to the company behind its technology.
Hippocratic AI, a US-based start-up, says a unique and bespoke algorithm used to monitor and tend to patients can be created by focusing on healthcare-specific data.
Munjal Shah, the company founder, said the new nurses could provide relief to under pressure NHS hospitals by ensuring its wards are “super-staffed” with both human and AI expertise.
“We’re building a large language model for healthcare, not for diagnoses, not for drug discovery, we really want to focus on all the different problems of in-service healthcare delivery,” Mr Shah told The Telegraph.
“We want to avoid diagnosis, the name of the company is Hippocratic AI after the Hippocratic Oath and our tagline is do no harm. We really feel that generative AI has a lot of potential in healthcare, but it should not be used for diagnoses, it’s just far too dangerous.
“In the UK, much like in the US, there is a massive staffing shortage. Why don’t we have enough nurses for all the people with chronic diseases?”
The data is specific to healthcare and runs through the vast processors of cloud computing giant Amazon Web Services which provides the technological heft to churn out the chatbot.
System can be tailored to perform specific roles
The technology is a chatbot-based system which can be tailored to perform the roles of specific healthcare experts.
For example, it can perform the niche of a fully qualified chronic care nurse, a pre-operative nurse and a genetic counsellor with an IQ of 130, a high score according to most intelligence measuring scales.
Hippocratic AI claimed its large language model passed more than 100 professional tests and certifications and outperformed ChatGPT in 92 per cent of cases.
Mr Shah says it costs less than $1 an hour to use the device and could therefore be used to give more people easier and more comprehensive access to healthcare advice.
For example, a chatbot pharmacist could help assess a new medication, while a chatbot geriatrician could ensure elderly patients are being cared for correctly.
“This is what I call ‘super staffing’,” Mr Shah said. “The opportunity here is not to try and save some money on the people we have, we don’t have enough nurses to begin with and we need to keep every nurse that we already have.
“But we need 10 times, 100 times the number [of staff].”
Voice-controlled version on the cards
The technology is text-based but the company is working on creating a voice-controlled version which a person could talk to in real time.
Mr Shah said this development would be more persuasive, natural and effective at delivering healthcare amid an increase in virtual consultations and check-ups.
“The thing we’re working on the most is giving [the technology] the ability to talk and to do speech recognition and speech synthesis over the phone,” he said.
“And that’s literally what we’re building because we know seniors are about 20 per cent of the population, but there’s 70 per cent of the patients and they like to talk on the phone and frankly, the phone is more persuasive.
“Voice is the original user interface and I think there is a really unique opportunity here to have a voice renaissance.
“People had to learn to write and use a computer. We forced people out of their natural communication style to these because of efficiencies and scale, not because they’re better.”
The company works with healthcare experts to help finetune the AI algorithm, but it is also working with Hollywood actors, Mr Shah said, to make the model as engaging, friendly and conversational as possible.
Not only will this help convince people to stick to a medicine plan, but it will also help keep lonely individuals company, another potential use of the technology.
AWS is taking on Google, OpenAI, Microsoft and others in the field of generative AI, and is working with a range of companies.
Alongside the chatbot doctor of Hippocratic AI, it is also underpinning BenevolentAI, a company which is focusing on making new drugs and used generative AI to discover that the anti-arthritis drug barcinitib may help treat Covid.
Tehsin Syed, general manager of Health AI, AWS, said generative AI, the process of a computer being trained on a vast amount of data and creating unique responses to questions based on this learning, has “enormous transformative potential for the healthcare industry”.
“With generative AI, healthcare organisations can leverage massive amounts of complex data to capture and present knowledge in more advanced ways.”
Another use of AI in healthcare is using it to automatically take notes without needing the doctor to manually transcribe them.
HealthScribe, for example, was launched this week by AWS and can be used to record, transcribe and fill in paperwork on its own. Babylon, the private healthcare app, is also following this approach.
“A great example of how we are bringing the power of generative AI to healthcare is this week’s launch of AWS HealthScribe, a new service that enables preliminary clinical documentation to be automatically created from patient-clinician conversations,” said Mr Syed.
“These conversations are the foundation for care, and we’re excited to help save clinicians time through the use of speech recognition and generative AI.
“We are taking generative AI and foundation models out of the realm of research and experiments, giving every company access so they can create positive experiences for their users.”