Microsoft Research has released a podcast episode featuring Jaime Teevan, Microsoft Chief Scientist and Technical Fellow, in conversation with three researchers behind the New Future of Work Report 2025: Jenna Butler, Jake Hofman, and Rebecca Janssen. The episode is part of the Ideas series, which focuses on the beliefs that animate Microsoft Research work. According to the episode description, the conversation covers AI adoption, AI’s impact on work, the intentionality required to create a future where people flourish, and the question of whether AI is better understood as a tool or a collaborator — and why that framing distinction matters.
The New Future of Work initiative has been running since 2020, when Microsoft researchers came together to make sense of the pandemic’s impact on work practices. The first report was published in 2021, and the effort has continued annually since. The 2025 report involved over 50 authors from across Microsoft and external institutions worldwide.
How the researchers describe the report and its purpose
Jenna Butler, who has been involved since the report began, frames its purpose as giving people agency in a moment when technology feels like something happening to them. “With any technology we introduce to society, that’s a sociotechnical shift,” she says in the episode. “How people perceive it, use it, what they want to do with it, what they’re willing to pay for — all these things matter.” The report, she describes, aims to make visible what current research shows, and to help people understand how their own behaviors and views are shaping the technology.
Butler’s research background is in software engineering productivity and, currently, AI’s specific impact on that domain. She notes that her original training was in bioinformatics and cancer research, and that she is drawn to multidisciplinary fields because the problems facing work today do not yield to single-discipline analysis.
Jake Hofman has been involved since 2023, with work focused on AI and cognition at Microsoft Research New York City. He co-leads a workstream called Thinking and Learning with AI, or TALA, with researcher Richard Banks. He joined as a section editor for the 2025 report. “I know how widely read and impactful the report is,” he says, describing the report as a valuable opportunity to present research from Microsoft and beyond with a coherent viewpoint.
Rebecca Janssen joined Microsoft full time in October 2024, having followed the New Future of Work reports from outside the organization during her PhD, where she studied AI’s impacts on work and society from an economics perspective. She describes the variety of topics and disciplinary perspectives in the report as one of its defining qualities.
The core tension: efficiency versus the future worth wanting
Jake Hofman draws what may be the episode’s sharpest distinction. “It’s easy for us to say, let’s get everyone to adopt and let’s boost efficiency. Let’s make everything really quick. But I don’t think that that’s actually the future, like, we want to live in.” That statement is not anti-AI — it is a pushback against treating adoption rate and speed as the primary metrics for success.
The implication is that optimizing for short-term productivity gains without asking what kind of work environment results is a design failure, not a design success. The report, as described in the episode, is an attempt to keep that broader question alive amid pressure to move fast.
Jenna Butler reinforces the intentionality theme. “It is not predetermined. The future of work is actively being built by us, by consumers.” That framing appears in the main report summary as well and is presented as a corrective to both techno-pessimist and techno-optimist framings — neither “AI will destroy jobs” nor “AI will fix everything” is treated as a given. The outcomes depend on choices being made now.
Benchmarking against the past as a mistake
Rebecca Janssen introduces a critique of how AI capabilities are currently evaluated: “We keep benchmarking against the past. So what can AI do, or can AI do what we already do? And I think this is, like, a mistake or maybe only the first step.” The argument is that measuring AI by its ability to replicate existing human work sets the ceiling too low. The more important question is what kinds of work become possible that were not feasible before — a question the research community has not yet answered because the answers require the technology to mature further and for people to develop new practices around it.
Teevan frames the current moment as a “really big shift in how digital technology can support people getting things done” — not a generational-scale change every year, but a structural transition that is happening over a longer arc. The podcast episode serves as a companion to the written report, offering the researchers’ interpretive framework for the findings in their own words.
The episode raises the tool-versus-collaborator question explicitly, though the excerpt does not resolve it in detail. The framing is treated as consequential rather than semantic: how AI is positioned — as a tool someone uses or a collaborator someone works with — shapes the expectations, trust calibration, and accountability structures that people bring to their interactions with it.