In today’s rapidly evolving technical landscape, artificial intelligence (AI) has emerged as a transformative force across industries. However, AI’s potential to revolutionize regulatory functions and drive strategic innovation remains largely untapped, especially in the life science industry. This article explores how organizations can leverage AI to not only streamline compliance processes but also to uncover new opportunities, mitigate risks, and gain a competitive edge over competitors. By embracing AI-powered solutions, companies can turn regulatory challenges into catalysts for breakthrough innovations and strategic growth.

Fundamentals of AI

AI is like giving computers the ability to learn and think in ways similar to humans, but often much faster. Imagine a tireless kid who can quickly go through vast amounts of information, spot patterns, and make predictions based on the available information. Being intelligent enough, AI can not only follow existing rules but also adapt and learn new functions, much like how we learn new things as a kid. From recognizing speech to even driving cars, AI is becoming an increasingly powerful tool to handle complex tasks, freeing up humans to focus on more creative and strategic work.

AI is not a new word. As an academic discipline, it can be traced back to Alan Turing. Researchers have been working on it for almost a century, and in the last three years, AI has become one of the most critical topics in everyone’s life. What triggers this dramatic change? Why has AI created so much buzz recently? We believe the key reason is that we are entering the era of generative AI. Here, let’s introduce two concepts: discriminative AI and generative AI. Discriminative AI, which we can also call traditional AI, is a type of algorithm that analyzes patterns and makes predictions by classifying new input into predefined categories. We have been using this type of AI for years without knowing it is AI, like facial recognition in our smartphones and spam filtering in email services. Discriminative AI takes input and gives classification answers, such as spam or non-spam emails. Generative AI is another type of AI algorithm that aims to create new content based on inputs, such as text, video, and music. Researchers have also been working on it for years, and ChatGPT is the symbol that generative AI models are mature and intelligent enough to produce meaningful content. The generative ability not only enables a wide range of new applications but also provides a more intuitive way for people to feel how intelligent those AI models are.

Get Your Organization AI-Ready

After learning about the power of new AI technologies, the next question we want to ask is, “Can we introduce AI to our organization?” At NyquistAI, we always consider the following two elements when introducing AI: historical data and success criteria.

The most beautiful feature of AI technology is its ability to learn new skills. We need to provide historical data for AI to learn from and give clearly defined success criteria to determine if our AI learns new skills. This is just like reading through the textbook and then taking an exam to decide whether or not we have learned those new skills. Historical data at your organization is the textbook, and the success criteria is the exam. If the answers to historical data and success criteria are “yes,” there is a potential for applying AI technology to your organization to boost efficiency.

Every life science company wants to embrace AI; however, only a few know how to think about and adopt AI. There are several common misunderstandings about AI that we need to avoid. First, AI can never be a wizard wand that solves every problem, especially cross-functional issues. AI can be your best assistant to save time, but it is still an assistant. We still need a human-centered partnership model of humans and AI tools working together for better results. It always requires humans to check results from AI models and make the final decision. AI is your tireless assistant, not an enemy. Second, we need to understand the limits of general AI tools. General tools like ChatGPT can work fantastically on general topics, but they lack the domain expertise and knowledge to answer specific questions in your daily job. It requires re-training AI models with companies’ data to meet specific needs. Without high-quality data, there won’t be meaningful deliverables from AI models.

AI’s Transformation of Regulatory Affairs

The integration of AI is not just about adopting new tools; it’s about fundamentally redefining the way regulatory affairs professionals approach their work. This transformation is ushering in an era where the regulatory function becomes more agile, efficient, and insightful than ever before.

AI has the potential to revolutionize regulatory teams by automating routine tasks such as data entry, document review, and post-market surveillance. These tasks, which have traditionally been time-consuming and prone to human error, can now be handled with precision and speed, freeing up regulatory professionals to focus on higher-level strategic initiatives. This shift allows RA teams to engage in more meaningful work, such as shaping regulatory strategies and anticipating future compliance challenges.

One of the most exciting aspects of AI in regulatory affairs is its ability to provide real-time insights through predictive analytics. By analyzing vast amounts of data, AI can enable RA teams to anticipate regulatory trends and align their strategies accordingly. This proactive approach not only mitigates risks but also positions organizations to capitalize on new opportunities within the regulatory environment.

One of my RA friends shared an interesting story about leveraging NyquistAI to identify competitors. The FDA has approved more than 170,000 devices and is approving almost 100 new devices each week. It is impossible to keep track of all new devices and identify competitors. Using NyquistAI to read through all documents on newly approved devices, she can get real-time insights and contribute critical insights to senior leadership meetings.

Dos and Don’ts for Using AI in Regulatory Affairs

As the adoption of AI in regulatory affairs continues to grow, it is essential for RA professionals to navigate this new terrain with care. Here are some “dos and don’ts” to ensure a successful integration of AI into your regulatory processes:

Dos:

• Invest in High-Quality Data Sources: The foundation of any successful AI initiative is high-quality data. AI systems rely on accurate, relevant, and up-to-date data to function effectively. By investing in robust data sources, such as government databases or historical submission data, organizations can ensure that their AI tools deliver reliable insights.
• Start Small with Pilot Projects: Before fully committing to AI, begin with small-scale pilot projects that target specific regulatory challenges. These pilot projects provide an opportunity to build confidence and trust in AI’s capabilities while allowing teams to learn and adapt to the technology in a controlled environment.
• Focus on Augmentation, Not Replacement: AI should be viewed as a tool that enhances human capabilities rather than replacing them entirely. Use AI to automate straightforward reviews, allowing human reviewers to focus on more complex cases that require nuanced judgment and strategic decision-making
• Prioritize Continuous Learning: The AI landscape is rapidly evolving, and staying informed about the latest advancements is crucial. Encourage your RA team to engage in ongoing education and training to stay updated on the latest AI tools and techniques, ensuring they can effectively integrate new AI developments into their workflow.

Don’ts:

• Don’t Rely Solely on AI for Critical Decisions: While AI can provide valuable insights, it should not be the sole decision-maker in critical regulatory matters. A thorough human review is essential in regulatory affairs to ensure that the context and potential risks are understood before a final decision is made.
• Don’t Overlook the Ethical and Legal Implications: The use of AI in regulatory affairs brings with it significant ethical and legal considerations, such as data privacy, algorithmic bias, and compliance with regulatory requirements. It is important to proactively plan how these issues will be addressed to avoid potential legal repercussions.
• Don’t Underestimate the Change Management Required: Implementing AI in RA processes involves more than just technical integration; it requires careful management of the cultural and organizational changes that come with it. Engage stakeholders early in the process, and provide the necessary training and support to facilitate a smooth transition.
• Don’t Ignore the Need for Customization: Off-the-shelf AI solutions may not fully meet the unique needs of your RA processes. Collaborate with AI vendors to customize solutions that align with your specific regulatory requirements and organizational goals, ensuring that the technology works for you, not the other way around.

Authors: Megan Kane, Director of Regulatory Operations at VelseraVelsera, and Michelle Wu, Co-founder & CEO at NyquistAI