On Friday, Elon Musk’s company Neuralink introduced the world to three pigs who seemed indistinguishable. Yet Gertrude, who was both shy and stubborn, had a secret: Two months prior, her brain had been implanted with Neuralink’s newest version of a brain-computer interface, or BCI. She looked just like another pig with no such device and a third who had had a similar device in and then removed. As Gertrude walked around doing pig things, viewers saw a display of her real-time brain activity. “If the device is lasting in the pig, as it lasted in there for two months and going strong, then that’s a good sign the device is robust for people,” Musk said.
Though the live event was filled with theatrical claims and outlandish promises, like being able to “store your memories as a backup and restore” them later, realistic advances in BCIs are definitely happening right now at Neuralink and elsewhere. Decades of work on BCIs could enable profound new treatments for patients with neurological injuries and conditions, such as helping patients with paralysis to control the mouse of a computer or move a prosthetic limb just by thinking.
The technology itself seems quite impressive and could be headed for clinical trials soon, which is exactly why we should take time to reflect on its goals. While Musk was emphatic that safety would be paramount—hence Neuralink showing off Gertrude and her compatriot whose device had been safely removed—the main issues addressed at Friday’s conference were physical harm and digital security. These are important hurdles for BCI development, but so are ethical concerns, which Musk mostly avoided during the event. One question left unanswered was: Who and what is Neuralink’s BCI really for, and how does that change where the technology is headed? While Neuralink made multiple appeals to helping people with disabilities, recent statements from Musk and his team are beginning to make these appeals look more like a thinly veiled front for transhumanist dreams.
Neuralink is one of multiple companies developing BCI devices. Neuralink’s approach has been to create both a new type of electrode—a neural “lace” or “thread”—and a surgical robot to “sew” the BCI into the brain. Moving from animal testing to human trials will require regulatory approval, and according to Musk, the Food and Drug Administration has given Neuralink a “breakthrough device” designation—meaning it can receive a faster review process. Friday’s event demonstrated progress in both the device and the robotic surgeon since last summer, and clinical trials are on the horizon, but it remains uncertain if the company will meet its initial goal of starting human trials in 2020.
Notably, Musk began the event by stating that Neuralink’s goal is to “solve important brain and spine problems,” and the team’s neurosurgeon later noted that clinical trials would be for people with spinal cord injuries. Neuralink’s 2019 concept paper suggests the same. While the initial goal of clinical use would be to help individuals with paralysis operate computers or machines, Musk suggested that Neuralink’s BCI could eventually be used to restore the use of limbs after a spinal cord injury. And the event included claims that the BCI could one day assist with conditions ranging from blindness to hearing loss.
However, comments from Musk and his team during and before the recent event were by no means limited to assisting people with disabilities. Musk suggested that not only could Neuralink’s BCI assist people with hearing or vision impairments; it could also allow users to hear frequencies and see wavelengths beyond normal human limits or to stream music right into their brains. In addition to hope for controlling pain and restoring function lost due to aging, the Neuralink team expressed ambitions to pursue applications of their BCI in gaming, “telepathy” for faster-than-speech communication, and to “save and replay memories” or send memories “into a new body or into a robot body.” In an MIT Technology Review profile of a Neuralink engineer released in July, the employee is quoted as saying “augmentation” for even healthy individuals “is an obvious result” of their work and Musk’s vision.
Musk also continues to cite his fears about artificial intelligence overtaking humans as reason to develop BCIs. He has made it clear that a long-term goal for Neuralink is to allow humans to “merge” with A.I., to make sure that humans can keep up with A.I. As he tweeted in July, “If you can’t beat em, join em: Neuralink mission statement.” And in an interview this summer, Musk claimed A.I. could become “vastly smarter than humans … less than five years from now,” evoking a sense of urgency to the goal of human enhancement. During Friday’s event, Musk reiterated that using BCI technology to find “A.I. symbiosis” “might be the most important thing that a device like this achieves.”
These comments make it difficult to figure out Neuralink’s goals: Is it to treat patients with neurological conditions and injuries or to pioneer human enhancement? If it’s both, Musk’s comments have made it clear he is more interested in augmenting healthy people, while treating patients along the way. And Neuralink isn’t the only BCI company with this dual purpose. For example, Facebook likes to talk about its work on BCIs as a way “to help patients with neurological damage speak again,” which is truly a noble cause. Yet, the company also has substantial interest in getting many of its users access to these wearable devices, not just those with disabilities, to increase data collection on users.
But who and what BCIs are being developed for matters. It changes whose interests are in the driver’s seat and whose are sidelined. Even without placing judgement on using BCIs for human enhancement, pursuing this goal with aggressive deadlines may create problems by leaving communities with disabilities behind and not considering the well-being of other users (beyond avoiding physical harm from the implant).
If therapeutic and enhancement BCIs are developed in tandem, how will we protect the interests of people with disabilities once they represent a minority of the marketplace? This group will likely be the first to test the technology and provide the data needed to improve it, but focusing on their needs could become less profitable as nonmedical BCI use takes off. Then, will assisting people with disabilities remain a focal point of Neuralink and others’ business models? As the percentage of BCI users who are patients or have disabilities falls, will companies continue to prioritize their needs over the interests of consumers?
Another question is: Who will own the enormous amount of “neurodata” generated by therapeutic BCIs, and what will be done with it? Could data from BCI patients with disabilities be used to help develop tools for human enhancement? Would they get a say? And just like with pacemakers, data from BCIs could likely be turned over to law enforcement without user consent, which could lead to extra surveillance for medical and nonmedical BCI users alike. Musk committed to making Neuralink’s device affordable at the August event, though getting more data from patient and consumer use alike will ultimately benefit Neuralink’s ability to develop its technology (to say nothing of how nonmedical data could be sold and monetized).
Neuralink and other BCI developers still have time to address these issues, even if their long-term goal is augmentation. Companies must speak with and—more importantly—listen to the patient groups and communities with disabilities they have pledged to help. If BCIs are really being developed to help those groups, then their needs and voices should influence where the technology is headed right now. And communities with disabilities can’t be treated as a monolith, either—cochlear implants should be a reminder that whether to use neurotechnology is a complex question, and individuals will differ on whether a device suits their desires.
Given the FDA’s role in approving BCIs, it should work with the industry and stakeholder groups to make sure BCI companies continue to update, maintain, and provide services to patients and users with disabilities—even after they start offering BCIs to consumers. And since HIPAA privacy rules only apply to health data, it might be time for new rules that protect neural data regardless of whether that information has a health purpose, which has already been done for genomic data.
Ultimately, BCIs like Neuralink’s look like powerful new tools that could create a future without neurological disease and injury, but also forever change the limits of human potential. Whose voices are heard and what risks we as a society are signing up for need to be addressed now.