When it comes to drug discovery, artificial intelligence (AI) is poised to change the game for the biotech and life sciences industries. This is equally true when it comes to the use of AI in the clinic. But as developers, programmers, and entrepreneurs have already begun to observe, the integration of AI in the healthcare and life sciences (HLS) industries is not as straightforward a task as some might assume.
Whether it is the adoption of technology, collection of data, mitigating AI hallucinations, or integration of new tech into established systems, the brave new world of AI is still being built and it won’t be for another few years before we understand what the world’s AI future will look like.
To discuss these topics, Thinkhat’s Chief Scientific Officer, and founding member, Dr. Sam Perli, hosted the AI-Driven Frontiers in Next-Generation Therapeutics session at the 2025 Precision Medicine World Conference in Silicon Valley.
“What AI has really done is give us a way to discover drugs in a non-traditional fashion,” began Dr. Perli. “We as scientists and engineers are used to hypothesis-driven experiments and with AI, that is changing. AI is generating the hypotheses—AI is even short-circuiting the hypothesis generation to identify the drugs themselves.”
“The classic analogy I always love to give is that back in the day, Newton sat under a tree and watched an apple fall and thought, What made the apple fall is what makes the planets go around,” Dr. Perli explained. “But these days you don't need this kind of hypothesis-based understanding because you can have a camera sitting in front of the apple, the apple falls down 1000 times, the camera takes pictures, and the camera now can predict when the apple will fall—it might even figure out F=ma.”
And indeed, AI models that are being built to observe and map the world are becoming more and more sophisticated by the day, thus opening the door for not only expedited drug discovery, but also improved diagnostics, the reduction of administrative burden (both in healthcare and outside of it), and beyond.
So what developments in the HLS industries are being most affected by AI, where is its adoption still lagging, and what does the future hold?
Issues still remain. As panelist, Shah Nawaz, Vice President of Technology & Digital Transformation at Regeneron explained: “With any technology, ecosystem-readiness is important.”
Just because you have the tech, doesn’t mean the ecosystem it is intended for can use it, Nawaz noted, and therefore there is a great deal of work that the AI industry has to do around its new tech to ensure proper adoption.
His sentiments were echoed by panelist David Rhew, M.D., Global Chief Medical Officer & VP of Healthcare at Microsoft. “We had another technology that everyone thought would transform the world,” Rhew recalls. “It would enable us to be able to have better outcomes, lower costs, and achieve greater satisfaction. That was the invention of the electronic health record (EHR).”
As Rhew recalled, even though the healthcare industry recognized that EHR technology had many capabilities, when the actual results were measured—whether they be quality of care, outcomes associated with that care, or patient safety—the industry saw huge variability.
“It was interesting,” Rhew continued, “the variability actually had nothing, or very little, to do with the technology itself. We saw the variability within systems that were implementing the exact same technology. And in fact, between vendors, there was very little difference. The greatest variability was within vendors. What we realized is that there is an association and a need for us to be able to think about how to implement AI specifically into workflows so that they lead to the outcome we want.”
Dr. Perli agreed, noting that the disconnect between AI developers and those using the technology down the line is something the industry is still trying to understand how to address.
“AI is still very new,” added Nawaz. “Of course, for us sitting in Silicon Valley, the tech is close to mature, but the moment you start to apply this in a business setting, we have to realize that, at least in our case in biopharma, we are still 10 to 15 years behind in overall adoption. Our ecosystem is simply not there. But you still need to make the best out of it. You still need to get people interested. You still need to identify those meaningful use cases that will truly aid value now.”
Dr. Perli also noted that even within the AI developer community, there is not yet a consensus around what degree industries should fundamentally utilize the technology.
“If building AI is one piece of the puzzle, it's important to note how we use it to drive applications that benefit humanity,” he said. “I think implementing AI comes with its own challenges. Some of us view AI as a tool, while some of us want to really unleash the full potential of AI—not only using it as a tool, but leveraging it to enable us in making decisions and offloading some of that burden. But then, that also comes with this idea as to how much we trust AI, and obviously the safety and the regulations surrounding it.”
None of these issues are small, but Dr. Perli’s conversation with the panelists exemplified just how seriously the industry is taking these issues and working to solve them. As such, it is not surprising that despite the work to be done, the future looks bright in the eyes of panelists
“I don't know how many of you have lived through the biopharma life cycle,” said Nawaz. “The average number of years to get the medicine out the door is still about 15—and every year it goes up. Which is ironic given the increase in technology.”
Perhaps it is the higher level at which life sciences researchers are working, or the regulatory hoops that companies have to jump through, or even improving access to more diverse patient populations for clinical trials, but Nawaz's observation is correct. The average development time for drugs and therapeutics is among the longest for investors and undoubtedly contributes to the tumultuous investment landscape that HLS entrepreneurs have to navigate.
As such, the introduction of AI signals what many hope is a beacon of light in choppy drug development waters—offering shorter development cycles, better management of clinical trial patients, and assistance when navigating regulatory pathways.
“One of the things I'm most excited about is AI applied outside of what we typically see to solve problems that we've never thought possible,” said Rhew.
“With these technological advancements there is the possibility to generate medicine,” concluded Nawaz. “That's a very bullish statement, but I do see a world where we will be in a position to possibly generate a drug. We'll still need humans, but we’re already seeing the tide change. We're seeing little pockets of the workflow being augmented, we’re seeing entire ecosystems evolve. Biotech and pharma adoption will come. I don't know if it will be three, four, or five years, but it will come, and I think we're just gearing up for that.”
In conclusion, Dr. Perli explained Thinkhat’s position in the developing AI world.
“I do feel like we will generate medicine, and that's the thesis of what we're doing at Thinkhat as well,” he said. “It's Computation Overcoming Disease, and—one thing I'm personally excited about—it is addressing the aging population and the aging epidemic.”
“We want older people to become healthier—and ideally younger,” he continued. “And we do think that it is possible biologically because older people give birth to younger people. So that’s one aspect of how and why we work with AI: to see how we can make that into a process, how we can make that into a therapy. I do feel like in five years from now, we're going to see some breakthrough in this space.”
Meet the only platform your finance team needs.
Powered by Thinkhat
