Forecasting AI blog





Brain Machine interface, spectacular advances

Brain-machine interface, spectacular advances

By G. H.


August 6, 2019





The brain-machine interface is no longer a fiction. Within a few days, Facebook and Neuralink reported major advances in this area. Review of details.

Sometimes there are events that require you to review your plans. In recent days, there have been many new developments in the field of brain-machine interface. Long time ago, it was just an inaccessible dream but two companies have made particular progress on this subject, enough in any case to stop and discuss about them. Present in all circles of Artificial Intelligence with his companies, Elon Musk, one of the founders of Neuralink, attacked the first and made promising announcements on July 16th, before Facebook followed suit a few days later. Mark Zukerberg's company also issued a press release reporting a publication by researchers at the University of California at San Francisco, which it supports financially, in the scientific journal Nature.

And the progress is purely impressive. Enthusiastic for some. Worrying for others who fear a loss of human control over the machine or, more simply, the end of private life or almost. It must be said that Facebook would apparently not be far from being able to read in your mind. It is certainly necessary to qualify because the study was conducted on three patients with short closed or semi-open questions on which AI was not even 100% successful. But we will come back to this later, because Neuralink had indeed drawn first!

Neuralink unveils itself

Neuralink was a startup that had never publicly communicated about its projects since its creation in July 2016. But three years later, we now know more about his intentions, which are commendable. "This has the potential to solve several brain-related diseases. The idea is to understand and treat brain disorders, preserve and enhance your own brain and create a well-aligned future," said Elon Musk during the conference on July 16th. To do this, the South African intends to connect your brain directly to any terminal using a chip. First, to allow people with disabilities to be able to control their machines directly by their thoughts. Then to achieve the perfect marriage between the human being and the machines in order not only to maintain total control over them, but also to improve the brain and physical capacities of the former. The famous myth of the augmented man.

The chip that Elon Musk would like to implant in the human brain for the brain-machine interface. (Credit: Neuralink)
Good talker, you might say? But he didn't wait for you and first tests on rats and a monkey were carried out! The publication of a white paper on the margins of the presentation reveals much more about the level of research achieved and the experiences gained to get there than just viewing the conference. As for the pictures of the rat with threads implanted in its cerebral cortex and a usb-c port directly in the skull, they certainly have an effect and need no comment...

Neuralink, your brain connected by USB-C to your machines?

To put it simply, Neuralink is now able to implant electrodes into the brains of rats using a neurosurgical robot of its own invention at a rate of "six wires (192 electrodes) per minute". A prodigious speed, and all this while avoiding blood vessels. The rest is a little more "artisanal" and less glamorous because the implantation of the electrodes requires piercing the skull and the chips are then connected to the machines by a USB-C cable. We have probably known more futuristic.... In any case, the businessman even plans to carry out the first tests on human beings in 2020. Of course, promises are only binding on those who receive them, but still.

A few days later, under the title: "Imagining a new interface: Hands-free communication without saying a word", Facebook reported a publication in the scientific journal Nature that made a big splash. Here again, it is the responsibility of the human being's brain and, more particularly, of his mind. Scientists at the University of California conducted a study, funded by the American firm, on the development of a brain-machine interface capable of detecting the words that a person is about to pronounce through the capture of waves emitted by the brain, in the language region more precisely.

Facebook wants to decode your thoughts

Three patients without speech impairment and undergoing surgery unrelated to the study volunteered to have electrodes implanted and participate in the experiment. The brain-machine interface had to be able to predict the question asked to the user by the simple activity of the brain and the answer with or without the specific context of the question. The questions were: "which musical instrument do you dislike hearing?"; "from 0 to 10, how comfortable are you?"; "how is your room currently?" What were the results? "We decode produced and perceived utterances with accuracy rates as high as 61% and 76%, respectively (chance is 7% and 20%). Contextual integration of decoded question likelihoods significantly improves answer decoding. These results demonstrate real-time decoding of speech in an interactive, conversational setting, which has important implications for patients who are unable to communicate," the researchers revealed in the abstract of the study. In the near future, these same researchers hope to decode brain waves at a speed of 100 words per minute, with a vocabulary of 1000 words and an error rate of 17%.

How about tomorrow, you connect to Facebook with your thoughts?
It goes without saying that Facebook's leaders and communicators welcomed these advances but also imagined the future and the applications that could result from them. "Being able to recognize even a handful of imaginary commands, such as "home", "select" and "delete", would offer entirely new ways to interact with current VR systems - and future AR glasses," the press release said on their blog. And, why not later imagine logging into your facebook account, scrolling through the news feed, liking, sharing and even commenting by the simple action of your thought?

The red line

Ethical and moral questions therefore arise when legislation does not currently provide any framework. How to protect the data stored in our brain? What about these private companies that fund and benefit from university research? For what purpose? Some scientists are already warning. This is the case of Nita Faharany, Professor of Neuroethics at Duke University and interviewed by the MIT Technology Review: “To me the brain is the one safe place for freedom of thought, of fantasies, and for dissent,” [...] We’re getting close to crossing the final frontier of privacy in the absence of any protections whatsoever." Marcello Ienca, a researcher at the ETH Zurich, said: "Brain data is information-rich and privacy sensitive, it’s a reasonable concern. [...] Privacy policies implemented at Facebook are clearly insufficient."