In an industry racing toward autonomous AI systems that can work independently for hours, one woman is betting nearly half a billion dollars that we’re asking the wrong question. Instead of “How can we make AI work without humans?” Andi Peng and her co-founders at humans& are asking: “How can we make AI that helps humans work better together?”
It’s a distinction that might sound subtle, but it could reshape how we think about artificial intelligence entirely. And for professional women navigating careers in an AI-transformed workplace, the answer to this question matters more than ever.
The Massive Bet on Human-Centric AI
On January 20, 2026, humans& announced a staggering $480 million seed round at a $4.48 billion valuation. For a company just three months old with no product in market, the numbers are extraordinary. But perhaps more extraordinary is what the company promises to build: AI systems designed not to replace human workers but to amplify human collaboration.
The timing is deliberate. As major AI labs pursue increasingly autonomous systems that can execute complex tasks with minimal human oversight, humans& is charting a different course. Their mission centers on creating AI that serves as connective tissue between people, strengthening organizations and communities rather than displacing the humans within them.
The vision has attracted backing from some of tech’s most influential players: Nvidia, Google Ventures, Laurene Powell Jobs’s Emerson Collective, and dozens of other major investors. That level of support suggests the market sees potential in an alternative approach to AI development.
The Woman Behind the Mission
At the heart of humans& is co-founder Andi Peng, whose career trajectory reads like a masterclass in staying true to your values while building technical expertise. After graduating from Yale, where she studied philosophy alongside cognitive science, Peng spent two years as an AI Resident at Microsoft Research. She then pursued her PhD at MIT CSAIL, working with professors Jacob Andreas and Julie Shah in the Embodied Intelligence Group.
Her research focused on building AI agents that learn continuously from and with humans. The work explored fundamental questions about how to align algorithmic systems with human needs, designing novel learning frameworks that center the end user. It’s research that would prove foundational to her later work.
Most recently, Peng worked at Anthropic as a researcher on reinforcement learning and post-training for Claude 3.5 through 4.5. She was instrumental in teaching these AI systems how to follow instructions and behave in desired ways. But as she watched the industry trend toward fully autonomous AI, she realized her vision diverged from that path.
“That was never my motivation. I think of machines and humans as complementary.”
Augmentation vs. Replacement: The Critical Distinction
The philosophical difference between humans& and other AI labs isn’t just academic—it has profound implications for how AI will reshape workplaces and careers. Most major AI companies are optimizing for autonomy: systems that can work independently, making decisions and completing tasks while humans sleep. It’s an approach with obvious economic appeal.
But Peng and her co-founders see a different possibility. Rather than training AI to work alone, they’re building systems designed to enhance human collaboration. Their vision includes AI that can ask meaningful questions, remember user preferences, and stay embedded in human workflows rather than replacing them.
CEO Eric Zelikman frames it as a matter of interaction design: “Chatbots are designed to answer questions. They’re not good at asking them.” The insight is that current AI doesn’t work hard enough to understand what users actually need. It waits passively for instructions rather than actively seeking to understand context, constraints, and goals.
For professional women, this distinction matters tremendously. An AI designed for collaboration could amplify uniquely human strengths rather than obscuring them.
For professional women, this distinction matters tremendously. An autonomous AI that handles tasks independently might seem efficient, but it can also sideline the human judgment, creativity, and relationship-building that remain irreplaceable. An AI designed for collaboration could amplify these uniquely human strengths rather than obscuring them.
The Technical Challenge
Building truly collaborative AI systems requires innovations across multiple technical frontiers. The humans& team is focusing on advances in long-horizon reinforcement learning, memory systems, and user understanding. They’re working on AI that can maintain context over extended interactions, understand individual preferences and working styles, and adapt to different organizational cultures.
The company’s vision includes software that facilitates seamless collaboration—think an AI-enhanced messaging platform where chatbots can request information from users, store it for later use, and help teams coordinate more effectively. Rather than completing work independently, the AI would serve as an intelligent coordinator and amplifier of human effort.
Peng’s background in human-robot interaction and social intelligence informs this approach. Her research at MIT explored how AI agents can infer human intentions and constraints by observing behavior, then make cooperative plans tailored to their human partners. It’s work that centers human needs rather than treating them as secondary to machine capabilities.
Why This Matters for Professional Women
As AI reshapes professional landscapes, women face unique considerations. Research consistently shows that women excel at relationship-building, collaborative problem-solving, and understanding complex social dynamics—skills that become more valuable, not less, as routine tasks become automated. But if AI systems are designed for autonomous operation, these strengths risk being undervalued.
A human-centric approach to AI could amplify these collaborative strengths rather than compete with them. AI that enhances team coordination, helps surface expertise across organizations, and facilitates better communication could make traditionally “feminine” leadership qualities more impactful. It’s a vision where AI augments human judgment rather than replacing it.
Peng herself represents what’s possible when technical expertise meets humanistic values. She didn’t just climb the traditional career ladder in AI—she questioned its direction and chose to build something different. Her journey from philosophy student to AI researcher to founder shows how deep technical knowledge can be wielded in service of human-centered goals.
The Skeptics Have a Point
Of course, a $4.48 billion valuation for a three-month-old company with no product invites skepticism. Some observers have called it evidence of a bubble in AI valuations. Others wonder whether humans& can actually build technology that’s fundamentally different from what larger, better-resourced labs are creating.
The challenge is real: every new AI lab promises a fresh approach until the realities of competition and compute costs kick in. Building AI systems that are genuinely different in kind—not just marketing—requires sustained commitment to principles that might not maximize short-term metrics.
But Peng’s track record suggests she’s playing the long game. She left a prestigious position at Anthropic precisely because she saw where that path led and chose differently. Her academic work consistently prioritized human needs over technical optimization. The question is whether a company backed by $480 million can maintain that focus as it scales.
What Comes Next
humans& is actively recruiting talent and building its team. The company has committed to contributing to open-source projects and academic research, suggesting a willingness to share insights rather than hoarding them. Their first products will likely emerge in the coming months, and the tech world will be watching closely to see if they can deliver on their ambitious promise.
For professional women thinking about AI’s impact on their careers, the humans& approach offers an optimistic alternative narrative. Rather than preparing for a future where AI replaces human workers, we might be headed toward one where AI makes human collaboration more powerful. Rather than competing with machines, we might partner with them.
Whether humans& succeeds or fails, Andi Peng has already made her mark by asking the right question. In a field that often treats human needs as constraints to work around, she’s building a company that puts them at the center. That alone makes her story worth following.
Because if we’re going to build AI systems that reshape how we work, we should probably ask: what kind of future do we actually want? And who better to answer that question than someone who’s spent years studying how humans and machines can truly work together?
* * *
About Andi Peng
Andi Peng is Co-Founder of humans&, a frontier AI lab dedicated to building AI systems that strengthen human collaboration. Previously, she worked at Anthropic on reinforcement learning and post-training for Claude 3.5-4.5, and completed her PhD at MIT CSAIL where she researched human-AI collaboration and embodied intelligence. A Yale graduate who studied philosophy and cognitive science, Peng has spent her career at the intersection of technical innovation and human-centered design. She has also worked at Facebook AI Research, Boston Dynamics AI Institute, and Microsoft Research.
Learn more: andipeng.com
About humans&
Founded in late 2025, humans& is a frontier AI lab focused on human-centric artificial intelligence. The company raised $480 million in seed funding at a $4.48 billion valuation from investors including Nvidia, Google Ventures, Emerson Collective, and others. Co-founders include Andi Peng (Anthropic), Eric Zelikman (xAI), Yuchen He (xAI), Georges Harik (Google employee #7), and Noah Goodman (Stanford). The team is building AI systems designed to enhance human collaboration rather than replace human workers.
