AI : Cortical Labs connects human neurons to a chip

AI has just crossed a strange frontier. Cortical Labs claims to have connected living human neurons to a silicon chip, with the ambition of creating a new form of computing, halfway between the classical machine and biological tissue. This is not just a gimmick. It’s a serious avenue towards more sustainable, more flexible computing, and potentially more bewildering than anything the industry has shown so far.

An AI no longer based solely on silicon

Cortical Labs no longer talks about purely software AI after the American restrictions on AI chip exports. The company highlights a system where real human brain cells are grown, connected to a chip, then stimulated to learn simple tasks. According to the story presented, these neurons were trained to play Doom in one week. The symbol is strong. It’s not about the game itself, but about the system’s adaptability.

What strikes is the change of logic. For years, modern AI has progressed thanks to more power, more data, and more data centers. Here, the idea is almost the opposite. Instead of stacking GPUs, the goal is to exploit the natural efficiency of living matter. The human brain remains, to this day, a brutal benchmark in terms of energy consumption and plasticity.

So this announcement should be read as a signal. AI is no longer just a battle between models, chips, and giant clouds. Another path is emerging. It remains experimental, limited, and still far from mass use. But it is already changing the debate.

Why this advance is so intriguing

The first point fueling interest around this biological AI is energy. According to information about the CL1 system, a rack of 30 units would remain within a very low electrical envelope compared to the huge infrastructures mobilized today to train and run AI models. That is where the topic gets serious. The energy cost of AI is now an industrial, economic, and political issue.

The second point is adaptability. An artificial neural network must be trained by computation. A living neural network learns differently. It reacts, reorganizes, stabilizes, or destabilizes according to stimulations. This difference can open the door to hybrid systems that are more economical and more flexible, capable of solving certain problems without copying the functioning of current architectures.

Finally, there is the accessibility of the model. With its idea of “Wetware as a Service,” Cortical Labs wants to offer remote access to this type of biological infrastructure. In other words, the company is trying to turn living neurons into a computing resource consumable like a cloud service. This is precisely what gives this announcement a sense of tipping point. Computing is no longer just being sold. Living matter organized to compute is being sold.

Between industrial promise and ethical discomfort

The discomfort does not come from a science-fiction fantasy. It comes from reality. When human cells from biological samples are used to perform tasks, even rudimentary ones, one question immediately arises: how far can one go without shifting the moral boundary? For now, we are talking about very simple systems. But the mere fact that this question already exists shows that the ground is sensitive.

One must also avoid getting carried away. Making neurons play Doom does not mean creating consciousness. It also does not mean that a machine “thinks” like a human. The risk here is not just technological. It’s narrative. The industry loves spectacular demonstrations. Yet a demonstration is not yet an industrial revolution.

That said, the concern is not absurd. If this form of Artificial Intelligence progresses, it could force regulators, researchers, and companies to rethink entire frameworks. Status of living matter, ownership of tissues, commercial use, liability in case of error: all of this will end up back on the table. And this time, the debate will not be theoretical.