The dream — or nightmare — of virtually instantaneous human-machine communications is a long-standing science fiction trope. Japanese aficionados have feasted on the concept for 30 years, since the publication of “Ghost in the Shell,” a story of police and political intrigue in a world in which human beings and computers are networked and the heroine is a cyborg — a human brain in the body of a robot.
The real world took a step closer toward that vision this week with the announcement by Neuralink that it has developed a robot that could implant threads deep inside the brain to facilitate communications between humans and machines. The prospect of a genuine link between the two remains a distant prospect, but we must as a civilization begin to grapple with the legal and ethical dimensions of this future and ensure that it proceeds in ways that advance, and not undermine, our interests as a species.
Neuralink announced it has developed a technique that allows it to insert a bundle of threads, each about the quarter of the size of a human hair, next to neurons in the brain. The threads are filled with sensors that capture information from a large number of brain cells and transmit that information, wirelessly, to a computer for analysis. Eventually, information will flow back and forth between computer and human. In one conceptualization, the threads would send information to a chip, which would transmit it to an external Bluetooth-like device behind the ear.
The idea is not purely theoretical. Neuralink demonstrated a system this week that reads information from 1,500 electrodes in the brain of a laboratory rat. That is 15 times better than current systems embedded in humans. The company also released a research paper that showed its scientists had completed 19 surgeries on animals, and that robots placed the threads accurately 87 percent of the time.
Neuralink is not alone in this effort. Several other companies, including Facebook, are working on human-machine interfaces. In keeping with one of the most prominent themes of science fiction, the U.S. military has been exploring implantable neural interfaces and one research program has yielded devices that allow quadriplegics to manipulate robotic arms to perform manual tasks like drinking. A Japanese company, Meltin MM, has developed a way to analyze the electric pulse that transmits commands from the brain to muscles and use it to manipulate prosthetic devices.
The military is not the only U.S. government entity engaged in this problem. The U.S. Food and Drug Administration earlier this year released draft regulatory guidelines for brain-computer communications. The guidelines are designed not only to ensure that this science is used properly, but to “promote innovation.”
Still, realization of this dream is a long way off. The Neuralink announcement focused on the success of the robot that inserts the threads; until now, it was impossible to implant, rapidly and precisely, sufficient numbers of sensors to collect data. Experts caution that success with laboratory animals is no guarantee of success with humans, the next step in experimentation. The company hopes to begin working with human subjects next year.
Company officials admitted that one of the reasons it went public was to lift the shroud of secrecy that had surrounded its work and begin to collaborate with other scientists. Neuralink plans to work with neurosurgeons at Stanford University and other institutions. There is hope that the techniques can combat serious brain disorders like Alzheimer’s, and that eventually they will become so safe and routinized that they are the equivalent of Lasik surgery and communications will be transmitted through a phone app.
Serial entrepreneur Elon Musk is one of Neuralink’s biggest backers, providing $100 million of its $158 million capitalization. Musk has warned against the threat of artificial intelligence, calling it “a fundamental risk to the existence of human civilization.” Musk, who was at this week’s announcement, said he sees Neuralink as a way to ensure that humans are not “left behind” by AI. Brain-computer interfaces will allow humans to achieve “a sort of symbiosis with artificial intelligence … to go along for the ride.”
That eventuality is, by most accounts, still well off. It should be noted, however, that “Ghost in the Shell” is set only a decade from now, in 2029. That does not leave much time to assess the implications of these developments and to understand the range of issues they raise. There are basic moral and ethical questions about the nature of humanity that demand answers, as well as practical concerns: Most fundamentally, can we ensure that these communications networks are safe and secure? There are no easy answers to these queries, no matter how enhanced our ability to think, calculate and reason.