WHO DARES IMAGINE THE NEAR FUTURE? ON SELF-TAUGHT AIs AND HACKABLE NEURAL IMPLANTS

This summer I started working on Indian sf writer Vandana Singh (see my post of 11 July on her short fiction) and I came to the conclusion that I really needed to do something to diminish my appalling ignorance of contemporary science. I mean something beyond reading science fiction… Just by chance I came across the daily newsletter offered by the Australian popular science magazine Cosmos and I signed up. For the last few months, then, I have been starting my working day by reading some of the (brief) articles referenced in their messages. It’s really very exciting.

The world looks different when you start paying attention to how scientists are fiercely arguing whether the fabled dark matter (the very fabric of the universe) exists or not, or when you are told that evolution might be actually be happening in just two generations and not as slowly as we believed. Last week I found myself voting for the best artistic image produced by scientists photographing brains: I didn’t know neurons could be that beautiful! Today I got enticed by an article about a marvellous finding of dinosaurs eggs in China… And, yes, of course, taking into account the horridly complicated political crisis that Catalonia is going through I can well say that science is giving me a different, healthier mental framework to cling on. The world looks in Cosmos far bigger and thrilling that that scary place portrayed in La Vanguardia or in The Guardian, with all the miseries of politics and economics.

This doesn’t mean that my daily intake of science is always tranquil. I am not at all, as you may guess, a technophobe and would even describe myself as a moderate post-humanist in favour of improving human existence with all the science we can use. Yet, I was very, very scared by a Cosmos article called “YouRobot: Neurotech may destroy your privacy and your rights”… Before I go into that, let me explain that while bland fiction about the life crises of old and young middle-class people, or about queens with dragons, or about surviving zombie hordes, occupies our attention, much better dramatic stories are going on in scientific research. You may have even missed the day when human history changed.

If you recall, IBM’s computer Deep Blue coolly beat world chess champion Gary Kasparov back in 1996-7 in a six-game chess match. Google’s program AlphaGo defeated go grandmaster Lee Sedol (the best player on Earth) by a 4-1 score back on 15 March 2016, a truly historic day. How’s AlphaGo (part of project DeepMind) different from Deep Blue? Brace yourself: AlphaGo is a self-teaching a.i. which learned to play the game of go better than any human being in just three days and after simply being fed the rules. In contrast, Deep Blue was fed a myriad matches, which it learned to process very fast. If you don’t see how AlphaGo makes a very deep difference in human History… you need to catch up. Urgently. See, to begin with https://cosmosmagazine.com/mathematics/all-systems-go-what-it-play-against-computer.

What really matters, I’m learning through Cosmos, is not at all what makes the front page. For instance, did you know that “116 founders of AI and robotics companies have called on the UN to ban lethal autonomous weapons”, that is to say, killer robots? (https://cosmosmagazine.com/technology/killer-robot-threat-must-be-faced-say-experts). Yes, James Cameron’s film Terminator (1984) is already happening. I learned in Richard Morgan’s novel Black Man (2013) that soldier robots are not in the end a very good idea, since they’re hackable. He speculated that genetically modified human beings might make the armies of the future, which sounds like replacing a terrifying nightmare with even a worse one… And it is happening.

If you’re not familiar with the word DARPA, then this would be the right time to check what they’re up to: https://www.darpa.mil/. I would say that the Defense Advanced Research Projects Agency is the closest thing we have to a classic mad doctor in Victor Frankenstein’s style but with all the power of the US dollar. And if you just thought ‘how about China?’, well, let me explain that China, Japan, Singapore and South Korea are now producing cutting edge a.i., both anthropomorphic and otherwise, without any shared ethical guidelines for research whatsoever (https://cosmosmagazine.com/physics/robots-bring-asia-into-the-ai-research-ethics-debate). In plain English: they’re building any kind of robot they can think of and nobody is checking on them. Nobody is checking on DARPA, either.

Before you think that I’m producing here a techno-advanced version of early 20th century ‘yellow peril’ let me finally focus on the scary story I mentioned earlier on (see https://cosmosmagazine.com/the-future/yourobot-neurotech-may-destroy-your-privacy-and-your-rights). If you have read William Gibson’s classic cyberpunk novel Neuromancer (1984) you must be familiar with the concept of jacking into the net: Gibson’s ‘cowboys’ are hackers provided with a cranial socket that allows them to link their brains to their computer terminals (or consoles) via a cable and a jack connector. Right, so it turns out that Elon Musk, Tesla’s charismatic founder, also runs a less well-known company, Neuralink (https://www.neuralink.com), devoted to building a brain/computer interface (BCI or ‘neural lace’). This was founded back in May 2016.

In principle, the idea of being able to think as fast as you computer sounds attractive, more or less, particularly if you still insist on playing go… I have always said that I would consider having my brain tampered with if this would give me better processing capacity, mad as this may sound. Yet, I had missed, as you will see, a crucial detail.

I have just learned from the corresponding Cosmos article that a group of scientists and specialists in ethics, calling themselves the Morningside group, have warned that the Declaration of Helsinki, the Belmont Report, and the Asilomar AI Cautionary Principles (documents about which I knew absolutely nothing) are not sufficient to prevent neural implants from being exposed to invasive manipulation and hacking. The implants now being used to correct motor deficit in sufferers of diseases like Parkinson’s, or to help paralysis victims to move objects, might be soon current–and capable of reading our thoughts. These are, just recall, electrical impulses that, sooner or later, will be decoded and controlled by powerful a.i. owned by corporations (and/or criminals)–a process to which Musk’s Neuralink is actively promoting.

The situation is so complex I can’t even begin to describe it but if you’re worried that the contents of your cell phone and your personal computer might be accessible to anyone (not just hackers but also bona fide, um, companies like Google), just think of the scary possibility that your brain might soon be equally unprotected. The Morningside group apparently believe that new ‘neurorights’ enforced through international legislation would protect, attention!!!!, our identity, agency and self-awareness. But even they acknowledge that “history indicates that profit hunting will often trump social responsibility in the corporate world”(see Rafael Yuste et al, “Four Ethical Priorities for Neurotechnologies”, Nature 8 November 2017, https://www.nature.com/news/four-ethical-priorities-for-neurotechnologies-and-ai-1.22960). This means that it’s game over for human beings are we are now. Privacy, consent, free will and identity, as they warn us, might be over soon. That is to say, the subject which is at the core of humanist thought.

The Morningsiders offer as a hopelessly optimistic solution teaching ethics to anyone involved in BCI technology. In this way, these persons would learn “to pursue advances and deploy strategies that are likely to contribute constructively to society, rather than to fracture it”. The group even proposes subjecting BCI sector workers to a new, specific Hippocratic Oath. Yet, it occurs to me that more than a few doctors must be in Elon Musk’s payroll already and the oath is not preventing them from opening up our brains to outside interference.

Perhaps, just perhaps, just as current teenagers think that privacy is a relative, overvalued concept which only worries Jurassic baby-boomers, the next generation will think nothing of having their brains directly linked to Google, Amazon or Facebook. Or their Tesla car…

Brave new world indeed.

I publish a new post every Tuesday (for updates follow @SaraMartinUAB). Comments are very welcome! Download the yearly volumes from: https://ddd.uab.cat/record/116328. My web: https://gent.uab.cat/saramartinalegre/

Leave a Reply

Your email address will not be published.