Artificial Intelligence (AI) is already transforming medicine in various ways, such as supporting doctors in making treatment decisions and improving the interpretation of imaging scans. Now, the curiously named AI tool ChatGPT and other next-generation chatbots appear poised to have a profound impact on both clinical care and research. At Massive Bio, we believe that AI chatbots could play an important role in the pursuit of our mission: Helping more people with cancer enroll in clinical trials of new oncology treatments.

ChatGPT is a rapidly rising star in the world of AI. This technology, developed by a San Francisco, CA-based company called OpenAI, has recently captured the imagination of scientists, scholars, journalists, and the simply curious. Other AI chatbots include Google Bard (which is still in development, but available for use) and Microsoft’s Bing Chat (which uses OpenAI’s GPT-4 platform). AI chatbots allow you to do something that would have seemed impossible not long ago: Have a human-like conversation with a computer. Ask an AI chatbot a question and, in mere seconds, you get a thorough and—in most cases—accurate response that’s composed in clear, plain language.

Most of us have encountered online chatbots by now, whether you were seeking technical support for a computer problem or had a question about a product you wanted to buy through an e-commerce website. First-generation chatbots can typically answer simple questions and direct consumers to resources where they can find help, but don’t do much else.

What sets apart ChatGPT and other AI chatbots is that they are trained with a staggering amount of data—essentially, all of the knowledge currently available on the Internet and more—which makes them experts on an incalculable variety of topics and capable of generating responses to most any reasonable request. Whether you ask where to buy a new bicycle near you or what caused the fall of the Roman empire, ChatGPT will tell you, in a conversational way that reads like an email from a friend who just happens to be the brainiest person in the world.

People are using ChatGPT for a dizzying array of purposes, ranging from scientists plumbing its knowledge to search for solutions to climate change to college students asking it to write term papers for them. Fortunately, nefarious uses of this technology appear to be outweighed by positive applications. Notably, a division of Microsoft Corp is currently working with OpenAI to learn what roles GPT-4 might play in healthcare and medicine. A recent commentary in the New England Journal of Medicine described some potential uses for this technology in patient care. For instance, an AI chatbot could “listen” to an office visit between a doctor and patient, then automatically draft medical notes for the encounter and order prescriptions or seek prior authorization for future treatments. After an office visit, a doctor might use an AI chatbot for a medical consultation to seek advice or guidance on the best treatment options for a patient.

Chatbots and Clinical Trials

At Massive Bio, we think this technology can help solve a persistent problem in medicine: low enrollment in clinical trials of new oncology treatments. Just 2 to 4 percent of people with cancer choose to participate in clinical trials. Among newly diagnosed patients, the numbers are even lower: only 0.1 percent enroll as part of initial therapy. This shortage of patients helps explain why about 80 percent of clinical trials fail to meet enrollment timelines and two-thirds of oncology trials are stopped before meeting their goals. The inability to attract sufficient numbers of volunteers to enroll in oncology trials slows the development of urgently needed new treatments.

The reasons for low enrollment in oncology clinical trials are complex, but a major part of the problem is that many patients find searching for an appropriate study to be too complex and daunting. Meanwhile, a cancer patient’s healthcare team may lack the time and resources to search for clinical trials. By combining the ability to process massive amounts of data rapidly with a user-friendly interface, ChatGPT and other AI chatbots could help minimize both problems.

“ChatGPT has the potential to accelerate the patient recruitment process for clinical trials in oncology,” says Cagatay Culcuoglu, chief technology officer, chief operating officer, and a co-founder of Massive Bio, which uses AI and concierge service to match people with cancer to oncology clinical trials. Culcuoglu argues that ChatGPT and related technologies could enhance these efforts.

For starters, ChatGPT could work seamlessly with Massive Bio’s SYNERGY-AI platform, which rapidly determines whether a patient is an appropriate candidate for a specific trial. Every clinical trial has precise guidelines (known as inclusion/exclusion criteria) that define the types of patients who are eligible to enroll. However, inclusion/exclusion criteria can be challenging for patients to understand and interpret. As a result, patients frequently seek to participate in trials that can’t accept them. That wastes precious time for patients and adds to the burden of trial coordinators.

ChatGPT could help eliminate this problem. “It can analyze large datasets and electronic health records to identify patients who meet specific eligibility criteria,” says Culcuoglu. “This can result in a more efficient and cost-effective recruitment process.”

As the technology matures, ChatGPT could one day be incorporated into the Massive Bio platform in a way that enhances clinical trial matching to an even greater degree, says Arturo Loaiza-Bonilla, MD, chief medical officer and co-founder of Massive Bio. “ChatGPT can interact with patients in a conversational manner, answering questions about clinical trials, eligibility criteria, and the overall clinical trial process,” says Loaiza-Bonilla. “This can provide patients with a more engaging and user-friendly experience.”

What’s more, ChatGPT uses a subfield of AI called machine learning that allows it to identify patterns within data, which allow it to make predictions that shape its responses to human input. “ChatGPT could learn and adapt based on patient feedback and interactions with the Massive Bio platform,” says Loaiza-Bonilla. “This could help improve the accuracy and relevance of clinical trial recommendations over time.”

Lack of awareness of and misconceptions about clinical trials are also barriers to participation. With their friendly, easy-to-use interfaces, AI chatbots can offer answers to common questions, provide details about specific trials, and help patients understand the potential benefits and risks of participation. “This can help to educate patients and increase their willingness to participate in clinical trials,” says Culcuoglu.

ChatGPT is capable of carrying on conversations with humans thanks to an AI capability called natural language processing (NLP). This technology is also what drives ChatGPT’s ability to translate text from one language to another. That includes translating medical information, treatment options, and other relevant materials, which can help international patients better understand their medical condition and treatment options, including clinical trials, adds Culcuoglu. It’s important to keep in mind that ChatGPT and other AI chatbots are in their infancy, and have experienced some growing pains, in a sense. Users have gleefully posted factual errors in responses generated by AI chatbots on social media. Worries that ChatGPT may be violating users’ privacy led Italy to ban the technology, at least temporarily. And, as with other AI applications (such as facial-recognition software), concerns about bias influencing the output of chatbots have arisen.

However, steps are being taken to address these issues. Meanwhile, the technology driving AI chatbots continues to improve. “At Massive Bio, we combine next-generation technology with patient-specific customization and customer service to find the right clinical trials for patients,” says Selin Kurnaz, Ph.D., the company’s chief executive officer and co-founder. “We’re hopeful that these emerging tools can help us achieve the goal that motivates us every day in a faster, more powerful and scaled way: Improving the lives of people with cancer.”