Artificial intelligence (AI) has incredible potential to enhance healthcare, from improving medical diagnosis and treatment to assisting surgeons at every stage of the surgical procedure, from preparation to completion.
Using machine learning and deep learning, algorithms can be trained to recognize specific medical conditions such as skin cancer melanoma. You can also use AI for image analysis to detect disease in photos with clean and documented datasets.
As a result, AI can help optimize the allocation of human and technical resources.
In addition, the use of large amounts of data by AI will make it possible to improve patient prognosis and treatment choices by adapting treatments to disease characteristics and individual idiosyncrasies.
Physician and medical consultant Dr. Harvey Castro cites the recent integration of Microsoft’s Azure OpenAI service with Epic’s electronic health record (EHR) software as evidence that generative AI is also playing an important role in healthcare. .
“One of the use cases is patient triage. AI is literally like a resident, where the doctor speaks, takes in all the information, and uses algorithmic understanding to initiate patient triage.” he says “If he has 100 patients in the waiting room, there’s a ton of information coming in so he can prioritize even if he’s not seeing them.”
“If you have 100 patients in the waiting room, you have a lot of information coming in, and you can prioritize even if you haven’t seen them.”
Dr. Harvey Castro, Physician and Medical Consultant
Castro added that it is important that any application of AI is meaningful and improves clinical care, rather than being introduced as a “shiny new tool” that does not help clinicians or patients.
He sees a future in which Large Language Models (LLMs) — large amounts of unlabeled text that help form the basis of neural networks used by AI — are written specifically for medical use.
“One of the problems with ChatGPT is that it wasn’t designed for healthcare,” Castro said. “To practice medicine, you need to be a correct LLM based on consistent, less hallucinogenic, referable and clear database data.”
The term “hallucination” refers to AI systems providing nonsensical or unrealistic responses or outputs.
In his view, the future of healthcare will be characterized by LLMs that evolve with more predictive analytics and can examine an individual’s genetic makeup, medical history and biomarkers.
Importance of regulation
Eric Le Quellenec, partner at Simmons & Simmons, AI and Healthcare, explains that regulation can ensure that AI is used in a way that respects fundamental rights and freedoms.
The proposed EU AI law, due to be adopted in 2023 and to come into force in 2025, will set out Europe’s first legal framework for this technology. A draft proposal was presented by the European Commission in April 2021 and is still under discussion.
However, AI regulations also apply to other European laws.
“First, the use of AI systems involving the processing of personal data is subject to the General Data Protection Regulation (GDPR),” he says.
Because health data is considered sensitive data and is used at scale, regulations mandate that a Data Protection Impact Assessment (DPIA) be performed.
“This is a risk mitigation approach that makes it easier to act beyond data protection and shipboard ethics,” added Le Quellenec, as did the Information Commissioner’s Office (ICO). data protection regulators have also released self-assessment factsheets. ) in England.
He added that the UNESCO Recommendation on the Ethics of Artificial Intelligence, published in November 2021, is also noteworthy.
“At the moment, these are all just ‘soft laws’, but we want to ensure that stakeholders have reliable data to use in AI processing and avoid many risks such as ethnic, sociological and economic bias. enough to make it,” he continues.
From Le Quellenec’s perspective, the proposed EU AI law, once adopted, should follow a risk-based approach and should avoid AI that creates unacceptable, high, low or minimal risk. use and establish a list of prohibited acts. Any AI system that is considered unacceptable for use.
“AI used in medicine is considered high risk,” explains Le Quellenec. “Before being placed on the European market, high-risk AI systems must obtain his CE certificate marking and be regulated.”
He believes that high-risk AI systems should be designed and developed in a way that ensures that their behavior is sufficiently transparent so that users can interpret the system’s output and use it appropriately. I’m here.
“All of this should give the public confidence and encourage the use of AI-related products,” said Le Quellenec. “In addition, human oversight aims to prevent or minimize risks to health, safety or fundamental rights that may arise when high-risk AI systems are used.”
This ensures that the results provided by AI systems and algorithms are used only as an aid and do not lead to loss of autonomy or impediment to medical practice.
Both Castro and Le Quelenec will speak on the subject of AI at the HIMSS European Health Conference and Exhibition in Lisbon from 7-9 June 2023.
