European Union Progress has been made on the long-awaited European Artificial Intelligence (AI) legislation. It was agreed to strengthen it in December. In this standard, the European body classifies medical devices containing this technology under the section High-risk AI systems. But what should the rules include regarding the health sector? Ignacio H. MedranoNeurologists and experts on this device summarize the essential elements into three aspects. First, it demands that AI not be classified as a medical device, as other rules do. “An algorithm built with AI machine learning It wouldn’t necessarily be a medical device, but that’s what has been decided.” Furthermore, it proposes a kind of pre-certification or ‘fast track’, as this would help encourage agility in development.
“In my opinion, I usa An error is made, which is then repeated United Kingdom, which aims to classify Artificial Intelligence as a medical device. That is to say, since I don’t have any regulator box to put it in, I just hook it up as if it were a catheter or electrocardiogram device so I can get it started. decision support algorithm, I have to do a clinical trial, and get the same seal from the FDA (US drug agency) that a machine has to get in a hospital. It doesn’t make any sense, an algorithm is information“, he explains. Moreover, this path ensures that it is neither sustainable nor scalable.
Therefore, the expert says Europe has the opportunity to “understand this better because it is coming later to the party”. This does not mean that AI algorithms are regulated: “Obviously, artificial intelligence algorithms have to be approved, monitored, and regulated, but they do not have to go through any process CE marking As if it were a physical device, that greatly delays innovation,” he says.
‘Fast Track’ to encourage speed
Furthermore, it proposes a type of installation ‘fast track’ Or pre-certification of companies or institutions to get things done faster. “It will be very important for the development of this technology and to have the opportunity to do it.”
To these two elements we add a third element related to the principle of data minimization European data protection law (GDPR, for its short name in English). As this regulation maintains, it advocates using the minimum possible data for each investigation. “The problem is that when you work with machine learning, Sometimes you find correlations between data that you weren’t previously aware of. So, really, the more data, the better,” explains Ignacio H. Medrano.
|
“AI algorithms have to be approved, monitored and regulated, but they also have to undergo CE marking”
|
That doesn’t mean it’s an excuse for companies to take ownership of the data, he says. “You can actually find relationships in the data that aren’t known until a machine starts looking at them.” Therefore, consider that a Artificial Intelligence and Health Law I need to keep this principle in mind to make an “exception” so that it does not represent a break.
Experts assure that this path is not free from difficulties, but it cannot be done by slowing down innovation, as this affects the technology reaching patients. “We have to provide guarantees,” he says, “but the problem with moving beyond that is that we don’t get the algorithm into the hands of people who really need it in time because they have a disease today, not in five years.” ”
Although it may include statements, data, or notes from health institutions or professionals, the information contained in medical writing is edited and prepared by journalists. We advise the reader to consult a health care professional with any health related questions.