AI startup training machine-learning models on living human skin tissue

Michael Polansky runs Outer Biosciences, a venture-backed startup that has spent years perfecting a radical approach to cosmetic research: keeping freshly harvested human skin alive outside the body for weeks at a time, then using AI to predict which untested chemical compounds might improve its function. This week, the CEO and co-founder broke years of near-total silence about the effort—and what the company has learned.

Polansky, who co-founded the firm in 2022 with chief scientist Kyung-Jin Jang, CTO Chris Hinojosa, and chief business officer Stanley King, is also Lady Gaga's longtime business and life partner. He remains executive director of the Parker Institute for Cancer Immunotherapy, a role he took on through his earlier work with venture capitalist Sean Parker.

How living tissue becomes a lab#

The speed of biological discovery has never kept pace with software innovation, Polansky argues, largely because there is no ethical way to test new compounds directly on humans. Scientists rely instead on animal models, cell cultures, and lab-grown organoids—all poor approximations of how human tissue actually behaves.

Outer Biosciences took a different path. The startup sources discarded surgical tissue, mostly from cosmetic procedures, through regulated biobanks and federally-funded repositories like the National Disease Research Interchange and the Cooperative Human Tissue Network. All samples arrive de-identified, stripped of donor names and contact information by the supplying organizations.

The company engineered a proprietary system that delivers nutrients and removes metabolic waste, keeping tissue alive for up to a month—far longer than the industry standard of just a few days. That extended window reveals biological processes that happen over weeks: collagen remodeling, pigmentation shifts, barrier repair, and inflammatory responses to injury.

AI meets living cells in a feedback loop#

The real innovation sits at the intersection of tissue and machine learning. An AI model generates predictions about which untested chemicals might improve specific skin functions. Outer Biosciences then exposes the living tissue to those candidates. The experimental results—success or failure—feed back into the model, refining the next round of guesses.

That closed loop is what Polansky describes as the actual product: not a single ingredient, but a faster discovery engine. Early on, the company mined scientific literature and studied natural compounds from extreme environments, producing a few leads over 18 months. Once AI was layered in, the pace accelerated dramatically. The startup now generates a viable candidate roughly every six weeks, with six leads currently in its pipeline.