Without a doubt, artificial intelligence (AI) is – and will continue – to reshape clinical research.
For example, earlier this month, exciting news broke that artificial intelligence (AI) invented two new potential antibiotics that could neutralize drug-resistance gonorrhea and methicillin-resistant Staphylococcus aureus (MRSA). The MIT team used generative AI (genAI) algorithms to interrogate 36 million compounds including those that either do not exist or have not yet been discovered to computationally screen them for antimicrobial properties.
The team used two different AI approaches to design new antibiotics.
First, they directed generative AI algorithms to design molecules based on a specific chemical fragment that showed antimicrobial activity, and second, they let the algorithms freely generate molecules, without having to include a specific fragment.”
In the words of MIT’s Professor James Collins,
We’re excited because we show that generative AI can be used to design completely new antibiotics. AI can enable us to come up with molecules, cheaply and quickly and in this way, expand our arsenal, and really give us a leg up in the battle of our wits against the genes of superbugs.”
As the antibiotic innovation shows, AI is moving from hype to reality. Study teams, sponsors, and regulators are both excited and cautious. Key concerns include data integrity and accuracy, ethical concerns and bias, interpretability and transparency, regulatory uncertainty and oversight, patient privacy and security, and clinical workflow integration. Especially in patient-centered research, there is a critical need to balance automation and accountability.
Where AI is Making a Demonstrable Impact Now
There are three primary use cases where AI currently is making a demonstrable impact in clinical research:
- Real-world evidence generation
- Regulatory submissions
- Biomarker discovery
Real-World Evidence Generation
AI is helping clinical trials to aggregate and analyze real-world data from sources such as EHRs, medical claims and patient registries. This allows for faster identification of eligible patients, improved trial design, and enhanced data accuracy, leading to more efficient and effective clinical trials. AI can also enhance post-market surveillance and safety signal detection.
Regulatory Submissions
AI is reducing the administrative burden of regulatory submission by automating tasks, improving data analysis and streamlining the generation of clinical study reports and summaries. Clinical trials also are leveraging AI for regulatory trend analysis and strategic planning.
Biomarker Discovery
AI significantly enhances biomarker discovery in clinical trials by enabling the analysis of massive, complex datasets and the identification of hidden patterns that traditional methods might miss. Clinical trials are using AI to analyze genomics, proteomics, and imaging datasets, providing a comprehensive view of patient biology and to identify potential biomarkers through pattern recognition and machine learning.
The Risk of AI: Moving Too Fast Without Guardrails
It’s tempting to want to deploy AI as quickly as possible in clinical trials. Perhaps you’re even feeling behind in AI deployment when you read about AI creating new antibiotics or a new AI software that is “twice as accurate” as professionals at reviewing the brain scans of stroke patients. Caution is warranted.
Ethical concerns such as bias, transparency and explainability are especially important in healthcare and clinical research. At the same time, many AI models, particularly complex ones, operate as black boxes, making it difficult to understand and explain how decisions are made. There are regulatory considerations and concerns including the need for rigorous validation to ensure accuracy and reliability, particularly when used to analyze data that will be submitted to regulatory agencies. Above all, sponsors must maintain auditability and control through robust data governance and the establishment of clear accountability.
Undoubtedly, AI is here to stay and will continue to change how we recruit, design and run studies. The future is bright, indeed. That said, it behooves all of us in clinical research to remember that real transformation requires more than just technology, it demands transparency, governance, and the right platform.