I spent much of this year debating the role of AI in each aspect of my life—working at Vega Health, conducting research, designing a product, figuring out the most effective way to study at school, and breaking down any subject where a Google search did not suffice. LLMs have enabled me to minimize busy work, distill large volumes of information into consumable insights, and reduce the barrier to knowledge and action. This has left room to exercise more creativity and expand the scope of what I previously considered achievable.
At the same time, knowing the capabilities of LLMs has triggered a race to optimize efficiency without a rulebook of constraints. Real standards for scientific integrity in AI use are still being defined and outpaced by the speed of LLM advancement, contributing to the AI slop penetrating both medical research and the healthcare industry. It has become clear that everyone has a different threshold for how to use AI with integrity guided by their own moral principles and increasing pressure to accomplish more tasks because of the efficiency gains from using AI.
At the start of this summer, my definition of using AI with integrity was confined to the act of reviewing LLM-generated code line by line to ensure it was doing exactly what was intended. After spending the summer with the Vega Health team, I realized how narrow that view was. Instead, integrity is about putting in the work to move from creating a product faster with AI to contributing to the world with AI.
Working with the implementation team, I witnessed how pivotal this team’s role is in fulfilling Vega Health’s mission of “turning AI investments into measurable results.” Considering that the value of an AI tool is only realized when it is adopted by end users, the implementation team works to promote adoption by building relationships early on. Throughout the discovery process, Vega Health connects with end users and leadership to ask questions, figure out what problem needs to be solved, determine and align with health system strategies, and teach end users the capabilities and limitations of solutions. Vega Health doesn’t just develop a solution and hand it off to its customers. Instead, this company partners with its customers to understand the problem they are trying to solve and co-design solutions addressing these problems.
Working at Vega Health has taught me that integrity in AI is beyond traditional scientific honesty. Integrity is fulfilling an obligation to use AI with the honest intention to make a positive impact. The integrity this team holds to ensure that Vega Health’s platform is not just a tool, but rather a solution, is what truly distinguishes them from the slew of tech vendors capitalizing on the LLM-driven reduction in the barrier to building AI products. Vega Health’s unique drive originates with the way every member of this team practices integrity in the individual work that they are doing through their dedication to excellence, willingness to ask for help, belief in the company mission, and respect for one another.
Building with integrity is a lesson I will take with me in my future career as a clinician. As questions float about whether AI could surpass doctors in diagnostic capability and perhaps replace them, I have come to take solace in the role that humans play in bridging the gap between creation and contribution. Throughout my numerous healthcare-related experiences, I have seen that the way to make impact has always been inherently human. Model evaluation has taught me that the value of a model is realized by determining the right problem that needs solving, communicating results effectively, and determining how solutions should be customized according to the workflow. Medical school has shown me that the value of a care plan is realized by building relationships with patients to promote transparency and determine how to better position patients to stick to that care plan. The value of health policy comes from aligning incentives between stakeholders, and the value of journalism is derived from asking the right questions to the right people.
From kitchen conversations to walking by cubicles, the people at Vega Health showed me the value of collaboration and humility in driving progress. I saw these traits in how our team asked each other for input, end users on their workflows and how to align our product with their goals, and advisors to see how Vega Health can continue to ensure it is moving towards their goal in a responsible way. Everything that I did at this company, whether it was working on an ROI calculator, creating workflow diagrams, designing an end user guide for how to use a product, and much more, stemmed from contributing to an actual need derived from a careful and ongoing discovery process.
Considering how LLMs can be used to automate tasks and improve the speed of creation, Vega Health taught me that the integrity to do due diligence, use our critical thinking skills, empathize with others, and strengthen our relationships with each other, is what will enable us to leverage AI to move from creation to contribution and truly make a positive difference with AI.
