Insight | Why AI's builders want medicine first: restoring lost function before augmentation
Neuralink, retinal prosthetics and deep brain stimulation researchers all start with patients who have lost something, such as movement, sight or impulse control, and only then talk about enhancement. This medical-first path shows how brain-machine interfaces will likely reach the public: through clinical need, not consumer upgrades.
The people building devices that plug into the nervous system start with patients who have lost something, and they give three shared reasons. The risk is only worth it when the benefit is large, regulators require safety proof first, and restoring a lost function is how scientists learn enough about a circuit to change it. Enhancement, if they mention it at all, comes after.
Matthew MacDougall, Neuralink's head neurosurgeon, says both the hype and the fear around the company are mostly wrong. The first product is meant to let quadriplegic people control a computer mouse and keyboard with their thoughts. At this stage they will not move their own limbs. He lays out his ambitions in order: reconnect the brain to the body, then over 10 to 20 years address problems like addiction and depression, and only in his children's lifetimes expand human cognition into AI. He would implant himself, but not yet, because early devices only do what his hands already can.
Stanford's E.J. Chichilnisky is building a retinal implant to restore sight to people whose photoreceptors have died. Today's implants give blobs and streaks of light, not natural vision, because they treat the retina like a grid of pixels. His slogan is "Make our instruments smart": record which cells a person has, calibrate, then stimulate each cell type correctly. He treats augmented vision as a spin-off of restoration, to be dialed up slowly.
Casey Halpern at the University of Pennsylvania uses deep brain stimulation for binge eating, OCD and addiction. His device detects craving activity and delivers a brief pulse to interrupt it. His conversation is entirely about treatment.
Mark Zuckerberg, who sells consumer hardware, makes the same prediction. He expects brain implants to start in medical cases, with people who have severe conditions and a very significant upside. His consumer bet is a wristband that reads signals the brain sends to the hand.
They disagree on two things. The first is whether a line exists between restoring and enhancing. Zuckerberg treats them as stages, while Chichilnisky says of that boundary "you can't even really draw a line", since even a basic vision device must choose whether to filter out infrared light. The second is tools. MacDougall thinks drugs will do more than implants for boosting plasticity, while Chichilnisky and Andrew Huberman see drugs as blunt next to precise stimulation.
Reading the sources together, the high-bandwidth implant path runs through medicine, while the consumer path runs through wearables that skip surgery and carry less information.
What to do with it: judge brain-interface news by which lost function a device restores and for whom. Watch clinical trials and registries, not product launches. Set realistic expectations for first-generation devices, and treat claims of memory or intelligence chips with suspicion. Expect consumer products to sit on the body, not in it.
For anyone considering these treatments, these are specialist decisions: talk with a neurologist, neurosurgeon, ophthalmologist or psychiatrist about what is actually approved and available for your condition.
