One-Sentence Takeaway

AI is most useful for Medical Affairs when it accelerates repeatable evidence, communication, and review work while keeping scientific judgment, patient context, and quality accountability with people.

Summary

Dan Chuparkoff opened by showing AI-generated video to make the current shift visible. His point was not that AI should replace experts. It was that the baseline for work is already changing, and people who use AI well will be able to work with more context, more speed, and more polish than people trying to do everything the old manual way.

The AutoCAD parking lot story introduced the keynote's central frame. AutoCAD did not become an architect. It helped architects do precise work faster and at a higher level of value. Dan used that same distinction for AI: the useful model is assistant, not automator. AI should help Medical Affairs teams move faster, communicate better, and review more thoroughly, but it should not own the judgment.

Dan then explained why AI can be both useful and wrong. Generative AI predicts likely next words from patterns. That makes it powerful for summarizing, drafting, comparing, translating, and finding signal in large bodies of information. It also means AI can be average, overconfident, outdated, or missing the context that only the human user has. The practical response is to ask for confidence, sources, assumptions, and review points before relying on important output.

For Edwards Lifesciences Medical Affairs, Dan connected AI to the work of accelerating evidence work while preserving rigor. He showed how AI can help draft FDA-facing narratives from grounded evidence, triage safety workflows, test trial designs with simulated patient cohorts, create stakeholder-ready explanations, translate statistical findings for different roles, answer expert questions from source-grounded information, surface deadline risk, filter signal from noise, and support quality review loops.

The closing message was the work pyramid: communication, process, and investigation sit at the bottom, while problem solving, decision making, and imagination sit at the top. AI is strongest at the bottom of that pyramid. The opportunity is to compress lower-level work so Medical Affairs professionals have more time for the human work that matters most: solving hard problems, making careful decisions, and imagining better futures for patients and teams.

Action Items

[ ] Ask an approved AI assistant the question Dan used: "What are nine ways AI will change the work of Medical Affairs scientists, biostatisticians, programmers, medical writers, trial designers, and safety reporters responsible for heart valve therapies in cardiovascular medical devices while accelerating evidence work, improving communication, and preserving rigor under deadline pressure?"

[ ] Pick one recurring evidence, writing, meeting, analysis, or review task and ask AI to improve the process, not replace the expert.

[ ] When AI gives an important answer, ask: "How sure are you? What sources did you use? What assumptions did you make? What should a Medical Affairs expert verify?"

[ ] Use AI as a reviewer on work you already created. Ask it to compare the work against clear criteria and identify how to make it stronger.

[ ] Identify one place where your team loses time translating technical evidence for different stakeholders, then test a small AI-assisted workflow for role-specific explanations.

[ ] Keep high-judgment work with people: scientific interpretation, patient-context tradeoffs, risk decisions, final quality review, and the future you want to build.

Key Ideas

AI is an assistant, not an automator.
The AutoCAD story showed the central distinction. A powerful tool can remove tedious effort and improve precision without removing the need for expert judgment.

The baseline has moved.
AI-generated content is getting good enough that work made without AI assistance may start to feel slower, thinner, or less complete. The practical question is where AI can improve today's work without creating unacceptable risk.

AI predicts, it does not know.
Generative AI works from probability. It can produce useful drafts, summaries, and suggestions, but it can also sound certain when it is not. Ask for confidence and source grounding when precision matters.

Average is not enough for expert work.
Dan described AI as a "B-minus student" at many things. That can be helpful for work outside your specialty, but Medical Affairs work often needs A-plus rigor. Use AI to accelerate supporting work and then apply expert review.

Human judgment comes from memories and hopes.
AI can summarize information, but it does not have your experience, ethics, patient context, team history, or goals. Those human inputs are part of real decision-making.

Use AI as a coach.
Instead of only asking AI to create a first draft, ask it to rank, critique, compare, and improve work against your standards. That helps people get better, not just faster.

Treat AI like an intern.
Give instructions, review the work, improve the instructions, and learn its capability ceiling. The review loop is the work.

The work pyramid should rebalance.
Teams spend huge amounts of time communicating, processing, and investigating. AI can help compress those lower layers so people have more time to solve, decide, and imagine.

Opportunities for Medical Affairs Professionals

• Move from periodic manual analysis to governed acceleration by using AI to keep evidence work moving as information changes.

• Draft FDA-facing narratives from grounded evidence, with Medical Affairs experts responsible for verification, interpretation, and final quality.

• Use real-time event triage to help safety workflows notice anomalies and surface the right work sooner.

• Test trial design ideas with simulated patient cohorts before real-world execution decisions are made.

• Convert protocol data, statistical findings, and technical details into stakeholder-ready explanations for different levels of expertise.

• Create role-specific briefs so collaborators receive the version of the evidence they can understand and use.

• Map expert questions to source-grounded answers that combine vetted internal context with appropriate external knowledge.

• Surface deadline risks before acceleration hits teams, especially when scientific rigor and timeline pressure collide.

• Filter signal from noise across incoming work so experts can spend less time sorting and more time deciding.

• Build quality review loops around AI output. The goal is not faster output alone. The goal is faster work that still earns trust.

Talk Flow

Digital emcee introduction

The introduction positioned Dan as someone who helps leaders turn complex technology into clear practical action, with a focus on what to hand to AI, what to keep human, and how to move faster without losing judgment.

AI-generated video and the new threshold

Dan opened with AI-generated video to show that AI-created material is becoming harder to distinguish from human-created material. The larger point was that the baseline for digital work is moving now.

AutoCAD and the copy-paste moment

Dan told the story of drawing parking lot lines by hand as a teenager, then seeing AutoCAD copy and paste precise lines in minutes. That moment became the keynote's main metaphor for technology assistance.

Assistant vs automator

Dan contrasted his view of AutoCAD as an architect's assistant with his boss's fear that AutoCAD would take architecture jobs. He applied the same distinction to AI: the tool should expand expert capability, not own the expertise.

Technology stair steps

Dan connected PCs, spreadsheets, the internet, mobile, cloud, data science, remote work, and generative AI as a sequence of work-changing tools. Each one eventually became part of how work gets done.

Crack open the AI pinata

Dan warned that many people are talking about AI without knowing what is inside it. Leaders need to understand which uses create real value and which create risk.

What GPT is doing

Dan unpacked generative, pre-trained, and transformer in plain language. The main point was that AI is generating likely output from patterns, not searching for perfect truth.

Once upon a time

The audience exercise showed that AI is consensus-driven. It often picks the most likely answer, which is useful for common patterns but weaker for ambiguity, novelty, or expert judgment.

Confidence and low-certainty answers

Dan explained that AI's confidence is usually hidden unless users ask for it. He encouraged the audience to ask how sure the AI is, where the answer came from, and what assumptions it made.

Glue in the pizza

The pizza example showed why AI can repeat bad information from the internet. AI can be useful without being a reliable final authority.

Autocomplete and human context

Dan used autocomplete as a familiar model for AI assistance. The tool suggests words, but the person still owns the message because the person has context, memory, judgment, and intent.

Memories and hopes

Dan argued that decisions are not driven by intelligence alone. People bring past experience, future goals, patient hopes, team context, ethics, and judgment to the work.

Medical Affairs AI question

Dan showed the audience the question he asked AI about Medical Affairs scientists, biostatisticians, programmers, medical writers, trial designers, and safety reporters responsible for heart valve therapies in cardiovascular medical devices. He encouraged attendees to ask the same type of question themselves.

Nine Medical Affairs transformations

Dan walked through nine AI shifts for Medical Affairs: FDA narrative drafts from grounded evidence, real-time safety triage, simulated patient cohorts, stakeholder-ready explanations, role-specific statistical briefs, source-grounded answers, deadline-risk surfacing, signal-from-noise filtering, and quality review loops.

AI before the big IT project

Dan noted that many large AI capabilities will arrive through internal or external software projects, but everyday AI use starts sooner. People can already use AI for writing, research, diagrams, inbox prioritization, planning, and translation.

Critique before drafting

Dan described a shift in his own AI usage: instead of asking AI to do the work first, he does the work and then asks AI to rank, critique, and improve it against his own standards. That turns AI into a coach.

AI as intern

Dan rejected the copilot metaphor and replaced it with the intern metaphor. An intern receives instructions, tries the work, gets reviewed, and improves through feedback. AI needs the same kind of review loop.

Accountability and the legal example

Dan used the lawyer case involving fabricated legal citations to show the cost of failing to verify AI output. The lesson for Medical Affairs was clear: AI can support research and writing, but the human expert remains accountable.

The hierarchy of work

Dan introduced six kinds of work: communicate, process, investigate, solve, decide, and imagine. AI is strongest at the first three. People should protect the last three.

Problem solving, decision making, and imagination

Dan explained why AI struggles with new problems, real decisions, and future imagination. Those areas depend on human context, values, experience, and the desired future.

Communication as leverage

Dan described the large share of work time spent in meetings, email, and chat. AI-supported notes, summaries, agendas, and retrieval can compress that communication load.

AI notes and AI song

Dan used Otter notes and an AI-generated song version of the keynote to show that AI can change how people capture, package, remember, and share information.

Translation and global collaboration

Dan showed AI translation as a practical communication shift. Language barriers are getting smaller, which matters for global teams, workforces, and patient populations.

Closing Message - The impossible world of more

Dan closed by naming the pressure of more information, more tasks, more stakeholder demands, and the same number of hours. The way forward is to use AI to compress lower-level work so people can spend more time solving hard problems, making better decisions, and imagining the future of Medical Affairs.