Haddonfield Memorial High School seniors worked with their lab director to publish a 15-page review of the benefits of AI in clinical trials in the February 28 issue of the Journal of Artificial Intelligence.
Meanwhile, the school district last month released a survey of students, staff, and parents about the potential impact of AI and how the district should take the issue. It found that the largest number of respondents, 54% of parents, supported the integration of AI in schools, followed by 44% of students and 32% of staff.
Mary Miao spent eight months writing the first draft of her review while attending to her regular school work and volunteering at Cooper University Hospital. This paper, “Applications of Artificial Intelligence in Clinical Research: A Narrative Review of Recent Advances and Challenges,” addresses the potential applications of AI in clinical trials, from participant selection to study follow-up.
Miao’s research searched for articles on AI over a five-year period. Her review describes how AI used in medical clinical trials can copy participants’ health data to increase a study’s sample size, scan social media to review trial results, and use chatbots to answer participants’ questions during trials.
Peng Peng Ma, head of the Institute for Environmental Molecular Medicine, who collaborated with Miao on the review, said that they both needed to learn about AI in order to write their papers and conduct their research. During the course of their research, they collected approximately 200 to 300 AI-related texts.
Miao explained that one of the reasons she decided to write the review was to counter fears about AI.
“…AI has a really bad reputation right now,” she said. “But I think it has much more potential to revolutionize medicine in particular, and I think it deserves recognition for that. Well, maybe not all AI is bad.”
Ma noted that recruiting research participants can be difficult, but AI could help make up for it.
“We can now use AI to mimic some patients,” he said. “You can base it on real-world data and pull data from that. That’s where AI can help. Just like AI can give you more candidates, more participants…this is one of the really great applications for clinical research.”
To reduce errors, Ma and Miao chose from three AI models. Although this review highlights the positive aspects of this technology in clinical trials, it also notes some drawbacks. Like humans, AI is also prone to biases based on the type of data used for training.
As an example, Ma and Miao observed that if an AI model primarily uses data from one ethnic group when selecting participants for a medical trial, participants from other ethnic backgrounds may not be included. It means excluding certain groups from the best medical care.
“It will further widen the health care disparities that already exist,” Maio explained. “Because if[the AI]is only trained for one specific group, all drug treatments will be tailored to that specific group. And people who aren’t in that group… won’t be able to get a drug that works for them, because the AI won’t have the information to create the drug they need.”
Part of the solution, she believes, is for those conducting clinical trials to learn more about the downsides and limitations of AI. Her review also points to data privacy concerns, as AI in clinical trials could mean exposing sensitive or personal patient data.
According to the district’s survey, one of the biggest concerns for responding teachers, staff, and students is that AI could be used for cheating, could be inaccurate, and that students could rely on it too much. Some comments in the survey were more critical of AI.
One said, “We need to allow students to do the work themselves before AI can enhance it.”
Others commented on the “huge costs” of AI and the impact on “water usage and electricity.”
Additionally, there were also positive comments.
“Not using AI is like not using the internet in 2000,” one respondent argued.
“AI is a useful tool at its best when humans have enough information to evaluate its output,” said another.
Although Ma doesn’t use AI for clinical research, he uses ChatGPT, Google’s AI, and Claude daily to summarize meetings and other administrative tasks. Miao uses AI to help her with her homework and explain math problems.
Despite the reconsideration, Miao said he would prefer to work with patients in family medicine rather than do further research. However, she would be interested in working with Ma again.
The Haddonfield School District recently released the results of an AI survey of students, staff, and parents.
