4 ways AI is reshaping discovery, health, work and responsibility

As AI becomes part of everyday life, its impact is increasingly showing up in concrete ways: in how scientists approach discovery, how doctors make decisions, how questions of accountability and inclusion are addressed and how work gets organized.

In the On Second Thought video series, futurist Sinead Bovell speaks with four Microsoft researchers and subject matter experts working directly in these areas. The conversations focus on what changes when AI is applied in real-world settings - what it makes possible, where its limits still are and how teams are navigating the tradeoffs that come with deploying these systems responsibly.

The series kicks off today, with new episodes coming Jan. 29, Feb. 12 and Feb. 26. Check out the trailer and highlights from each episode below.

Episode 1: How AI could make medicine more personal by enhancing human care

AI's most immediate value in healthcare is helping clinicians navigate the sheer complexity of modern medicine, says Jonathan Carlson, vice president and managing director of Microsoft Research Health Futures. No physician can track every specialty, study or guideline, Carlson says. But AI can add to the digital tools doctors already rely on and help them more quickly diagnose an ailment, bring together relevant medical knowledge and apply it to the specific patient.

"We're really diverse," he says. "All of us are different. And yet medicine has to operate off of averages. Yet none of us are average. Every one of us has our idiosyncrasies, right? And so the goal of personalized medicine is to do the right thing for you, to do the right thing for me."

The technology helps address the mismatch of treating deeply individual people with systems designed for populations, Carlson says. In oncology, for example, treatments only work for a fraction of cancer patients, with no clear way to predict who will respond.

AI can help by structuring and analyzing vast amounts of messy clinical data - written notes, PDFs, faxes, scans and more - so physicians can make more precise decisions.

AI agents can even take on different specialists' roles to synthesize data that helps doctors see the bigger picture, he says. But they won't replace human care.

"We as humans need human interaction, need human touch, need human judgment," he says. "Framing the question of will it be an AI or a doctor is just a false dichotomy. It will obviously be both."

Watch the trailer

Episode 2: Why building AI responsibly requires more voices, earlier

AI is reshaping society, but who gets to shape AI? For Hiwot Tesfaye, a technical advisor in Microsoft's Office of Responsible AI, the answer has to be: everyone. AI systems need input not only from technologists, Tesfaye says, but from linguists, social scientists and anthropologists, as well as from everyday users around the globe.

"As many opinions and perspectives as we can incorporate as early as possible in the design of an AI system and the training of an AI model, the more voices we can incorporate, the better," she says.

Tesfaye draws on her own experience growing up across Ethiopia, Sudan, Kenya and Uganda, where exposure to different cultures at a young age shaped her ability for "playing translator" between cultures and continues to inform how she approaches her work at Microsoft. That global lens also drives the Global Perspectives on Responsible AI Fellowship program she leads at Microsoft, which convenes AI experts from the Global South including policymakers, engineers and the creative community to better inform Microsoft's Responsible AI program and to help avoid blind spots as new systems and models are built.

The goal of the fellowship is for Microsoft to be in the seat of the student and have the fellows "help us understand what are the beneficial uses of AI that are manifesting your part of the world, but also, what are the misuses that you're seeing that we need to be aware of," she says. "Our technology touches every corner of the world, so we need to be aware of how these misuses are playing out."

Watch the trailer

Episode 3: When AI changes the pace - and structure - of scientific discovery

AI isn't simply speeding up lab tasks - AI agents are reshaping discovery by working alongside researchers through every stage of the scientific process, synthesizing vast amounts of existing research, making connections across disciplines, generating hypotheses and even helping to design and run experiments.

"We think of it in the terms of the entire scientific method," says John Link, a partner product manager for Microsoft Discovery, who focuses on chemistry and materials science. "It's the right AI agents, working hand-in-hand with a scientist throughout that process."

Microsoft looks for domains where new technologies can create outsized impact, Link says. Science stood out as one of those areas after the company used AI to help identify a new, more environmentally friendly data center coolant molecule in less than 10 days - work that would have taken years using traditional methods.

"It was our 'aha' moment that, hey, if we can do it, like, just imagine what this could do for all of our customers," Link says.

Looking ahead, Link envisions labs where "every entry-level scientist" arrives already supported by "a team of virtual postdocs." The promise, he says, is progress on global challenges like food insecurity and climate change.

"In the end, we need to solve the world's problems faster," he says. "We need to accelerate scientific discovery. And we think these are great tools to do it."

Watch the trailer

Episode 4: As AI democratizes expertise, skills matter more than titles

AI is leveling the playing field at work as people in every stage and role are collectively navigating the impact of the technology on their jobs, says Colette Stallbaumer, who is the general manager of the Microsoft 365 Copilot and Future of Work teams. As AI democratizes expertise, it's also flattening traditional hierarchies and changing how teams form and function inside organizations.

"One thing that we see is this dynamic of what we talk about in my team and describe as moving from the org chart to the work chart," Stallbaumer says. Instead of static roles and titles, employees increasingly come together around projects, then disband and regroup as priorities shift. "Teams are fluid, more agile," she says, adding that AI allows people to build expertise beyond a single discipline.

That means one of the most important mindsets for employees now is to be a lifelong learner, Stallbaumer says, adding that AI skilling is a business imperative.

"It can't be an afterthought," she says. "You're either going to disrupt yourself or be disrupted in this environment."

This reality is reshaping how companies hire and how careers grow, Stallbaumer says, with curiosity, adaptability and the ability to learn in real time becoming more valuable than years of experience alone.

This moment is uncertain, she says, but also full of possibility: "It's an incredible time in history for people to lean in and learn and be optimistic."

Watch the trailer

The full series will be available here.