Lab-grown human neurons learn to play DOOM: what happened and why it matters
First Pong, now DOOM: researchers are pushing living human neurons into roles we used to call purely digital. What began as simple game tests is becoming a proof of principle for biological computing. The leap raises real questions about how we manage this work.
Researchers in Australia at Cortical Labs reported that clusters of lab-grown human brain cells can be trained to interact with the classic first-person shooter Doom. The story reads like a science fiction pitch, but the methods and the numbers are concrete. This is not just a novelty demo anymore.
If one country starts genetically modifying its population, others won’t stand by and watch. It’ll set off a global arms race—not with missiles or drones, but with DNA to see who can engineer the smartest, most advanced generation of humans on the planet. – Steve Watson
The work builds on an earlier milestone where neurons learned to play Pong, and it now moves into 3D, dynamic environments. Cortical Labs describes clusters of roughly 800,000 to one million living human neurons arranged in a dish and connected so they can receive patterned electrical input from a computer. Those patterns are interpreted by the cells and translated into in-game actions.
Engineers converted Doom’s outputs into signals the neurons could register and then mapped particular firing patterns to actions like shooting or moving. The system observes which patterns yield success and reinforces those responses over time. That closed-loop learning is why observers say the cells appear to be adapting in near real time.
Cortical Labs chief scientific officer Brett Kagan explained in a video announcement: “So we showed that biological neurons could play the game Pong,” “This was a massive milestone because it demonstrated adaptive, real-time, goal-directed learning.” He and his team stress that Doom presented a harder, more realistic challenge.
Building on that 2022 achievement, the team tackled a tougher challenge. “Doom was much more complex,” Kagan added. “It’s 3D. It has enemies. It needs to explore, its an environment, and it’s hard.”
The platform, identified as the CL1 biological computer, permits remote access to these living cell clusters through an online interface. For now the neurons perform like beginners, but the pattern of improvement is clear and measurable. That trajectory is why the project is drawing attention beyond the lab demo community.
Cortical Labs chief technology officer David Hogan described how actions link to neural activity: “If the neurons fire in a specific pattern, the Doom guy shoots,” “If they fire in another pattern, he moves right.” Those literal mappings are how the team translates cell behavior into controllable outputs.
Back in 2022 the Pong demo grabbed headlines by showing biological systems could adapt to simple tasks, and now a more sophisticated test uses roughly 200,000 neurons interfaced with silicon to enable gameplay. The company and outside commentators call the CL1 a “code deployable biological computer” that has been shipped since last year. That language signals a move toward commercial and developer access, not just lab curiosity.
With that shift come immediate ethical and security questions. Advocates highlight potential in neuroscience, AI research, and robotics, while critics warn of slippery slopes toward transhumanist applications and dual use technologies. Conversations about governance, oversight, and transparency are accelerating as a result.
Imagine biological neural substrates contributing to autonomous systems, surveillance tools, or battlefield tech. Even if those scenarios feel speculative today, the technical building blocks are being assembled and tested. That makes public debate and clear safeguards essential.
Cortical Labs maintains the neurons are still learning basics and are far from expert-level control, but the development path is obvious. Who funds this work, what rules guide it, and how will misuse be prevented in areas like military AI or population-scale interventions are all open questions.
These experiments sit at the crossroads of biology, computing, and policy, and they demand attention from scientists, ethicists, and regulators. The lab demos are striking on their own merits, but the broader implications are what will determine how society responds.