Embrace AI in regulatory, but ensure humans are ‘in the loop’, experts say

ROTTERDAM, NETHERLANDS — A panel of experts at RAPS' European Digital Technology and Software Conference on Thursday urged regulatory professionals to embrace artificial intelligence (AI). They said the technology is already here, and while there are initial shortcomings, regulatory professionals can help adopt the technology for their organizations in a safe and responsible manner.

Maarten ter Mors, CTO of CREATE Regulatory Consultancy; advised the audience to start using AI if they’re not already. "Do not sit on the fence and wonder what to do … I think that phase has passed." he said. "AI is here and it's here to stay."

He added that society is adopting AI technology three to four times faster than it adopted earlier technologies like email and smartphones. However, he also noted that many AI implementations will not come from executives in boardrooms but from workers in the field.

"A bunch of senior VPs in the room, they're not going to set out the lines of what we're going to do in a field that's changing so quickly," said ter Mors. "Empower people to start experimenting and finding those use cases that really add value."

Janine Swarbrick, a partner at HGF Limited, noted the technology’s impact on intellectual property law, which tends to move very slowly as a profession. She said the technology can help summarize large stacks of documents and pinpoint key parts that may be useful to legal experts, saving the users a significant amount of time.

As a patent attorney and co-lead at her firm's medtech and digital health section, Swarbrick said she has seen a lot of impressive innovation using AI but cautioned that the technology is still evolving and needs a “human in the loop.”

Ter Mors echoed that AI is well-suited to processing large amounts of information and extracting highlights and summaries, though he has also seen many customers use it for more analytical tasks. He noted that technical dossiers have evolved over the past 15 years, and users can use AI to assess whether their product's intended purpose is consistent across all their documents. In that situation, he said it would be up to humans to use the technology to ensure that the AI outputs consistent and valid information.

"Those tasks can be sped up immensely, and in an innovation context, that means that regulatory [considerations] can be at the beginning of the product development and not just policing it at the end. They can be proactive," said ter Mors. "They can already be in the generation of the product, because ... we can do certain things a lot faster than before."

Swarbrick added that users should be cautious and triple-check everything the AI does. She highlighted cases where AI has hallucinated and invalidated people's work.

“I feel like everyone should kind of just have a growth mindset or a curiosity mindset, to say, ‘Actually, where could AI help our businesses?’” said Swarbrick. “I think at this point, one of its strengths is to go through all of that kind of mundane type of work.

“Even if it doesn't give you an answer, it can certainly give you some guidance or a starting point to then say, ‘You know, this is where we can focus our efforts,’” she added. “AI might not do your job for you, but it can help you balance out your workplace so that you can add your value in a meaningful way.”

Joseph-Richardson Larbi, director of medical device regulatory affairs at Celegence, stressed that, regardless of people's skepticism, AI is here to stay.

"We are in an environment where, unfortunately, you can't do anything about it," said Larbi. "The horse is already bought. AI is here. We just need to embrace it.

"We need to approach it with some curiosity ... we need to set boundaries," he added. "I'm sure we'll come to that in terms of the regulation side of things, and how we can use it safely to make sure it doesn't really breach the human way of living."

In terms of AI's potential in the regulatory field, Larbi said that he recently asked ChatGPT to draft a template to classify medical devices in accordance with the EU Medical Device Regulation (MDR) 2017/745 and was amazed at how accurate and thorough its response was.

"It mentioned the annex, and it gave me all the rules, and said, 'You can now add to this in this particular format," said Larbi. "I sat there in shock... I was like, 'Wow, AI is about to take our jobs.

"But [on the other hand] it's giving you more time to focus on other areas," he added. "If you look at it that way, you can embrace it."

Ter Mors argued that throughout history, when new technologies have arrived, people have worried that they would lose their jobs, but instead, they have adapted their jobs to accommodate the technology. Similarly, he said it's essential that people evolve with AI.

"The job will change, but the job will not be replaced," said ter Mors. "The regulatory community, people in this room actually have a head start over everyone else, because we are already used to dealing with risk classification in a medical device... We can also help organizations that are not so used to this sort of thinking to gauge the risks and benefits of AI and make sure that we put the human in the loop."

Embrace AI in regulatory, but ensure humans are ‘in the loop’, experts say

Related topics

Recommended content

Welcome to the new RAPS Digital Experience