Leen Kawas on AI and drug discovery: The ethical challenges that will define the next decade
[Disclaimer: This article is intended for US audiences]
In the biotechnology sector, artificial intelligence (AI) has emerged as a transformative force with profound implications for drug discovery and development. According to biotech leader Leen Kawas, the integration of AI with biological data marks a critical juncture that presents both unprecedented opportunities and complex ethical challenges.
Leen Kawas, an experienced biotech industry leader, highlights the benefits and challenges of integrating AI technology into the sector.
“AI enables us to bring a number of different data together to empower more accurate and comprehensive decision-making,” explains Leen Kawas, Managing General Partner at Propel Bio Partners, a Los Angeles-based venture capital firm supporting early-stage biotechnology companies. “This capability drives rapid depiction of multiple-level cellular processes, interactions, and reactions, enabling a better understanding of the relationship between expansive datasets and faster health-related discoveries.”
The Current State of AI in Drug Discovery
The conventional drug development cycle typically spans 10 to 15 years from initial discovery to market approval, with an estimated cost of $2.6 billion per successfully developed drug as of 2021. For every 5,000 compounds that enter preclinical testing, only five typically advance to clinical trials, with just one ultimately receiving approval.
AI technologies are now being deployed to address these inefficiencies. “We see a big surge in the number of companies that are trying to use AI and predictive modeling to accelerate drug development and discovery,” notes Leen Kawas. This acceleration stems from advancements in computational power and data science capabilities that have enhanced researchers’ ability to capture and process increasingly large and diverse datasets.
The integration of AI into pharmaceutical research functions is projected to shorten development cycles while better predicting which drug candidates are more likely to succeed. “AI-based start-ups have the promise of identifying and designing new drug candidates in a more efficient and effective way,” Kawas observes. “AI holds the potential to reduce timelines for drug discovery, improve predictions on clinical efficacy and safety, and can diversify drug pipelines without any bias from individual experience.”
Biomarkers: The Critical Connection Point
A key component in AI-driven drug discovery involves biomarkers—biological indicators that can be measured to assess normal biological processes, pathogenic processes, or responses to therapeutic interventions.
Leen Kawas highlights how biomarkers interact with AI to advance precision medicine: “Biomarkers may be an ideal solution, as they can identify the patient population that can benefit most from specific therapies, model response, and effects that can increase the success of potential therapies. As personalized medicine gradually comes into focus, each individual’s biomarker panel can help with patient selection and stratification.”
This integration of biomarkers with AI and machine learning has shown particular promise in addressing challenges in rare disease drug development. “Rare disease drug development attracts an increasingly large share of investor dollars,” Kawas explains. “However, drug developers have found it difficult to identify patients for drug research and treatment activities. Biomarkers can identify the patient population that can benefit most from specific therapies.”
Practical Applications Beyond Theory
The theoretical benefits of AI in drug discovery are increasingly translating into practical applications. Leen Kawas points to significant developments in several areas:
Diagnostic Performance Improvement
“Accurate diagnosis of medical problems is key to resolving them,” Kawas states. “Every year, close to 12 million Americans who seek outpatient medical care are misdiagnosed. Some diagnostic errors have resulted in life-threatening outcomes.”
AI-enhanced analysis of medical imaging and biological samples offers potential solutions to this problem. “AI can often analyze medical data more quickly and accurately than a clinician,” notes Kawas, though she cautions that “AI technology is currently not capable of completely replacing skilled radiologists.”
Personalized Treatment Advancement
The conventional approach of assuming uniform patient responses to medications is increasingly giving way to personalized treatment models supported by AI. “Patient-specific treatments and medications are increasingly becoming an option,” Kawas explains. “Healthcare providers are now monitoring certain treatments’ results via wearable sensors and trackers.”
AI-enhanced software can then analyze these treatment outcome patterns and recommend the most effective interventions according to each patient’s profile data. This approach promises to reduce complications and associated costs while improving outcomes.
Chronic Disease Prediction
Early detection of chronic diseases enables timely treatment and often better outcomes. “For each patient, obtaining lifestyle as well as clinical data helps to paint a picture of that patient’s health and risk factors,” Kawas observes. “On a macro scale, health experts can gather multiple patients’ data to predict a population’s emerging health problems.”
AI and machine learning can process information from electronic health records, social media, and online surveys—data that would be impractical to analyze manually. “These technologies can provide researchers and healthcare providers with a range of useful insights in a fraction of the time required for existing data-gathering and analysis methods,” says Kawas.
Ethical Challenges on the Horizon
While the potential benefits of AI in drug discovery are substantial, Leen Kawas identifies several critical ethical challenges that will shape development in this field over the coming decade:
Data Privacy and Security
As biotechnology increasingly relies on vast databases of patient information, protecting this data becomes paramount. “Biotech clinical trial databases typically contain patients’ personal identifiers, medical records, wearable-generated information, and even genetic sequences,” Kawas explains.
The expanding scale of these databases increases vulnerability to breaches. “An expanding patient database means more individuals’ personal identifiers and records are at risk from a data breach. The global average cost of each data breach is approximately $4 million.”
Leen Kawas emphasizes that biotech companies must implement stringent data protection measures: “Each applicable biotech must maintain secure servers and stringent access controls. Sensitive data encryption is also required.” Failure to do so risks not only financial penalties but also loss of patient trust, making future clinical trial recruitment more difficult.
Algorithmic Bias
AI systems are only as unbiased as the data used to train them and the people who design them. “When leaders lack all relevant information, they may postpone a decision until they have a clearer grasp of the situation,” Kawas notes, highlighting the importance of thorough validation before implementing AI-based solutions in healthcare.
Diverse development teams can help address this challenge. “Women in executive positions tend to cultivate more inclusive cultures,” Kawas observes. “This creates environments where team members feel comfortable expressing concerns about potential negative implications of technological applications.”
Regulatory Framework Adaptation
The rapid advancement of AI in healthcare creates challenges for regulatory bodies designed for traditional drug development paradigms. “The FDA’s existing drug development protocol was not designed for increasingly complex medical treatments,” Leen Kawas explains.
Creating appropriate regulatory frameworks that ensure safety while enabling innovation remains a crucial challenge. Leen Kawas advocates for “beneficial regulatory reform” that could simplify the drug approval cycle without compromising safety, potentially resulting in significant cost reductions.
Maintaining Human Oversight
As AI plays an increasingly prominent role in drug discovery and healthcare decision-making, maintaining appropriate human oversight becomes essential. “While supporting teams through uncertainty remains important, successful leaders also maintaina clear focus on strategic objectives,” Kawas states, emphasizing that technology should augment rather than replace human judgment.
This balance is particularly important when addressing complex ethical questions that arise during drug development. “Creating an environment where innovation flourishes requires establishing deep trust within teams,” Kawas notes. “This mindset shift from pure metrics to holistic development creates space for creative thinking and calculated risk-taking.”
The Path Forward
Looking ahead, Leen Kawas believes that the successful integration of AI into drug discovery will require thoughtful navigation of these ethical challenges. “Technology can lead to better tools for individualized and precision medicine. It allows us to make sense of the different factors that can make each individual or patient unique,” she states.
This human-centered approach to technological advancement will be essential for realizing AI’s full potential in biotechnology. “Using AI to have a holistic view of patients and individuals can lead to the discovery of new therapies or technologies that can help humans live healthier and better,” Kawas concludes.
By addressing data privacy concerns, mitigating algorithmic bias, adapting regulatory frameworks, and maintaining appropriate human oversight, the biotechnology industry can harness AI’s capabilities while minimizing potential harm. As Leen Kawas and other industry leaders guide this transformation, their ability to balance technological innovation with ethical considerations will shape the future of healthcare and drug discovery for decades to come.
ADVT.
This article is brought to you by Dr. Leen Kawas.
