Artificial intelligence (AI) technologies, like ChatGPT and Google Bard, are taking the world by storm. While most news coverage to date has focused on the potentially negative effects of these tools, it’s not all doom and gloom.

When used strategically, AI has the power to help businesses save time and money, improve data accuracy, and in the case of biopharmas, bring drugs and therapeutics to market sooner.

Since there’s so much hyperbole and misinformation out there, we wanted to take a closer look at what exactly AI is, its potential benefits for our industry and how you can use the technology to your advantage.

What is artificial intelligence?

Artificial intelligence uses computer systems and large data sets to teach machines how to perform functions typically done by humans. Movies like The Terminator depict AI as being sentient and all-powerful, but the reality is much less scary.

AI technologies may seem like they’re capable of thought, but this isn’t the case. It might be at some point in the future. But right now, any function an AI performs is the result of exposure to data and regular practice with the help of human-designed algorithms.

This explanation is admittedly somewhat technical. So, to put it very simply – AI is capable of analyzing and gaining insights from large swaths of data quicker than humans. It’s this ability that makes it exciting for biopharmas.

Potential Applications for AI in Clinical Trials

There are many potential applications of AI in clinical trials, several of which we have outlined below.

1) Data Analysis

Hundreds of thousands of data points are collected during clinical trials. While traditional statistical analysis provides important insight, it’s nowhere near as thorough as AI algorithms.

Modern AI technology can identify patterns and insights that humans might overlook or miss. You can even use it to analyze data from previous clinical trials and outside sources to amplify your understanding and make adjustments.

2) Protocol Design

Up to 90% of clinical trials for new drugs fail. This statistic is alarming, but AI may be the key to more positive outcomes.

AI can analyze large data sets and identify key factors that might affect study outcomes.

Considering that any hiccups during a trial require amendments to the protocol –– and in turn, weeks or months of delays –– using AI in the design phase could make a significant impact.

These insights can then be used to reduce the risk of failure and increase the likelihood of success.

3) Patient Recruitment and Selection

The recruitment period for industry-sponsored phase III clinical trials is 18 months, on average, and often longer.

Finding patients who qualify is a tedious and time-consuming process that requires lots of research and a thorough screening process. Even in the best circumstances, it can be challenging to identify and recruit a diverse patient base. Here too, AI offers benefits.

Many biopharmas now use AI algorithms to analyze electronic medical records and other types of patient data. This speeds up the selection process and makes it more equitable. It also makes it easier to choose patients who are likely to respond to the new treatment (predictive enrichment) and benefit from it (prognostic enrichment).

Here’s a real-world example: Researchers at Mount Sinai Medical Center in New York recently used a topological data analysis (TDA) to identify three subgroups of people with type 2 diabetes by analyzing electronic health records and genotype data.This analysis helped the team identify specific patients who might benefit from the medication being tested.

4) Patient Monitoring

Patient safety is a key concern for biopharmas, but even with the best screening processes in place, it can be difficult to predict negative outcomes.

AI algorithms can be used to detect adverse effects or potential safety concerns in real time. These insights can then be used to improve participant safety and lower the number of study dropouts.

When combined with wearable technologies, like smartwatches and clothing-embedded sensors, it’s possible to extend these safety benefits beyond the clinical setting. In other words, AI improves the safety of remote, hybrid, and decentralized clinical trials

Here’s another real-world example: several types of AI software have been developed to predict drug toxicity based on target information. This technology is still in the early stages, but researchers hope that it will eventually replace in vitro and animal models as the traditional clinical approach.

As you can see, AI offers the potential for improving efficiency, reducing costs, and enhancing patient safety. But it’s certainly not a replacement for the human element in research.

Clinical trials involve complex ethical and scientific considerations. Considerations that only humans can make. Therefore, AI should be used in conjunction with the systems and people involved in the clinical trial, not in lieu of them.

Things to Consider Before Implementing AI in Clinical Trials

Before you fully embrace AI, it’s essential to consider the potential risks of using such new technology. We aren’t going to argue for or against AI as a whole, but there are several areas of concern, including:

1) Ethical and Regulatory Considerations

AI technology offers lots of exciting potential, but there are plenty of ethical considerations, including how you access and utilize patient data. Currently, there’s very little regulation of AI tools in drug development. The FDA has developed a proposed regulatory framework, but nothing is set in stone. At this moment, there are a lot of gray areas and little guidance.

2) Privacy and Security Concerns

Patient privacy is strongly regulated in the U.S. So much so, that it can be hard for patients themselves to access their own medical data. Legislators and other decision-makers are considering safeguards to help regulate AI and data privacy. But the technology is so new, it will probably be years until there’s any solid movement. Regardless, not taking privacy and security seriously could lead to patient distrust, not to mention fines and penalties.

3) Over-dependence on AI and a Lack of ‘Human Touch’

AI technology makes it easier to interpret data and gain important insights into potential drugs, therapeutics, and patient populations but it isn’t a replacement for already established processes. If anything, AI should be used as a supplement rather than a substitute. Trial participants are humans and they deserve to be treated as such.

4) Bias and Fairness Issues

Some tech enthusiasts consider AI a panacea for societal problems like racism and sexism. Although it’s true that machines don’t judge, plenty of research has confirmed that biases are built into AI algorithms. For the technology to reach its full potential, these biases will need to be identified and weeded out.

AI in Biopharma – Takeaways

AI offers a variety of potential benefits for biopharmas, including improved data analysis, better protocol design, easier patient selection and recruitment, and improved safety. But even with all of these advantages, there are ethical concerns to consider, like privacy, security, and bias.

We’re on the cusp of something incredibly exciting. But with so much uncertainty, it’s essential our industry doesn’t run before it can walk. AI technology isn’t going anywhere. However, we must be strategic about how we use it for the sake of clinical research and humanity as a whole.