Michael DiMercurio and Kevin Warwick

I’ve been thinking a lot lately about artificial intelligence since hearing Professor Kevin Warwick’s interview on Red Ice Creations Radio. I’ve written about my initial thoughts on the Voice: http://hpanwo-voice.blogspot.com/2008/08/kevin-warwick-on-ai.html I also wrote a full-length article after watching a BBC Horizon film about the subject: http://hpanwo.blogspot.com/2008/06/ai-and-new-human.html . What Warwick has to say about AI is very interesting and in some ways he’s got his head screwed on; in others he’s deeply naïve. He can’t conceive how artificial intelligence and psychotronics could be abused by authorities and quite relishes the notion of a cashless society where there’s no need for keys to open doors because of the electronic implants in human bodies. He doesn’t understand the potential for abuse such technology has. My journey into the discovery of AI has become more intense since I’ve also been reading a series of novels by an American author called Michael DiMercurio (http://www.ussdevilfish.com/index01.htm). He used to serve in US Navy submarines and today writes thrillers on the subject which are very good stories, if somewhat specialized towards the niche market of the die-hard sub-fan. One of the most interesting aspects of his writing is his use of the concept of artificial intelligence. In his seventh novel Terminal Run he invents an unmanned submarine called “SNARC”, which has a genetically-engineered artificial biological brain running it. It is euphemistically termed a “carbon computer” while ordinary computers are called “silicon computers” to distinguish them. This brain has thoughts and feelings of its own just like any human. These carbon computers can't be treated like any previous electronic device. They can’t be developed through upgrading, they have to grow and learn just like a human neuronet does. They suffer from the "terrible two's" and the programmers find that the only way to cope with this is to give them something to vent their anger on. They wire them up to robots with weapons and let them loose on old cars and buildings. Once the AI's have destroyed a few things they feel better and can continue to develop. Part of the plot is that the scientist who created the SNARC, and who helps the villain Alexei Novskoyy hijack the robot submarine, doesn't want the sub to come to any harm because he's begun to see the artificial brain as a friend. The programmer can actually talk to the AI brain and have a conversation, and even have a relationship with it as a real, conscious entity! Alan Turing would be jumping up and down with delight! This raises an interesting question: If you could create artificial intelligence and therefore machines with thoughts and feelings just like living creatures, then you'd no longer be able to treat them as machines as you would a car or Hoover or toaster. You'd have to consider their feelings! There would be a whole new discipline of law: robot ethics! These machines would need legal rights like animals do under today's law. Isaac Asimov addresses this question in his Robot novels. One of them was made into a film, Bicentennial Man starring Robin Williams as a robot who is unintentionally created with artificial intelligence, and although his mind is not entirely the same as a human one, he is undeniably conscious and self-aware. The film follows his poignant struggle to be accepted for what he is in human society (http://www.imdb.com/title/tt0182789/). In Terminal Run, Noveskoyy’s mission is to use the SNARC to attack an American submarine. Now with the electronics we know, the world of “silicon computers”, this is simple. You simply reprogram the computer; and you can do so in any way you choose because the computer “brain” is not really a brain at all and the word is used by electronic engineers in a purely metaphorical sense. But SNARC is different; it is naïve and child-like, but questions and analyses completely independently. As the brain of an American submarine it feels uncomfortable about its new mission. Novskoyy is faced with the task of reasoning with the machine, persuading it that engaging the friendly submarine is a legitimate action. He spins the SNARC a yarn about the boat being taken over by mutinous murderers, but he cannot command it! This is the big dilemma with AI, you can’t have your cake and eat it! If you design a conscious and intelligent machine then don’t expect it to be the mindless slaves that conventional computers are. I’d go as far as to say that any computer without that analytical ability and the self-will to act on its conclusions cannot by definition be intelligent.

The most interesting, disturbing and amusing part of the book is when the AI designers put a carbon computer in a torpedo! In DiMercurio's early books, he uses metaphors of artificial intelligence in relation to conventional computers. His first novel Voyage of the Devilfish has in it high tech torpedoes called Mark 50 "Hullcrushers" which he describes as having the same level of intelligence as a Golden Retriever! Does this mean it would fetch a stick if you threw it (Could be dangerous with a torpedo!)? Would it sense if you came home after a bad day and sit beside you to comfort you? It's pretty obvious that in his early writing career, DiMercurio didn't really understand much about artificial intelligence because the Hullcrusher torpedoes when fired traveled straight towards their target and detonated without a second "thought". If the human weapons controller aboard the submarine sent a signal along its guidance wire to change course it would do so as surely as a car does when you turn the steering wheel. A Golden Retriever wouldn't do that! So really the Hullcrusher torpedoes were no more intelligent than a car steering wheel. This changed by the time he wrote Terminal Run. The designers of the new torpedo, called “Tigershark”, are faced with a conundrum: As I said above, an artificially intelligent computer is by nature, independently-thinking and cognoscent. One of the inherent instincts of conscious intelligence is self-preservation. But being an intelligent computer inside a torpedo means that to do your job you have to destroy yourself! You have to swim up to a target and blow yourself to pieces. In order to override this universal natural drive against self-destruction, the designers have to make it suicidal! This then causes the AI's to become psychotic! Once the Tigersharks are launched they immediately turn around and attack the ship or submarine which launched them! Such properties make them useless, as well as too dangerous, for anything except being launched from an aircraft. This I think could be the first case that “robot ethics” lawyers would need to table. Isn’t it cruel to create artificial intelligences like that? After all to give an animal a drug that would make it psychotic and suicidal would be out of the question (unfortunately this does go on illegally). In the book the hero, the submarine captain Michael Pacino, takes his sub out to sink the SNARC armed with the new Tigershark AI torpedoes. His crew at first think he's crazy, as does the reader, and wonder if he's on a suicide mission. But then all is revealed. He gets the submarine's medic to inject a sedative into the torpedo's artificial brain to render it unconscious. He then launches the torpedo in the general area of the SNARC and dashes away as fast as he can before the torpedo's brain "wakes up." When the brain does wake up the only thing it can hear around it is the SNARC so it goes for that and sinks it, leaving Pacino's own boat alone. In real life such weapons would probably never be developed. The whole problem with the AI weapons in DiMercurio's books is that they wouldn't just blindly obey the instructions of their human masters. They'd analyze and cogitate and ask questions. They might say: "What's in it for me?" and quickly reach the inevitable conclusion: "Not a lot! When I reach this enemy submarine and detonate my warhead I'll die! Perhaps I'd better not." Not very helpful in the middle of an undersea battle! The only alternative to that would be a torpedo like the Tigershark designers eventually developed: a psychotic torpedo with a hate-filled death-wish. This, as I’ve said, would be highly dubious, both practically and morally.

And this is just one example of the disadvantages of AI. Imagine an AI car that refused to take you to the shops because it got bored of just continually driving the same route over and over again! It might be on its way to pick you up from work and decide to take a trip to the seaside instead! You could only ask it to change its mind; you would not be able to command it. No futurist or sci-fi speculator has ever considered this: That maybe we should hang on to our good old-fashioned obedient and subservient silicon computers, even when the technology exists to make them obsolete!

Despite Kevin Warwick’s apparent lack of understanding of the Dark Side of AI; he also shows a remarkable open-mindedness in other aspects of it that are absent from so many others in his field. The majority of his peers take a reductionist view; that the development of AI and its inevitable back-engineered discoveries of the human brain will simply expose all spiritual thought as a delusion. But Warwick takes the opposite view, what you might call “AI-transcendental”. He points out that far from eliminating them, AI would augment many philosophical questions about the nature of mind and consciousness itself. For example: would AI machines have a soul? Seeing as we are their creators does that make humans into God? Would they continue to exist after they were shut down, ie "killed"? Is there an AI afterlife? Michael DiMercurio’s views on spirituality are very interesting because they develop and change enormously through his writing career. In his early books they’re completely absent and hardly discussed; in his fourth book Piranha Firing Point one of the characters, Pacino’s Uncle Dick Donchez, is dying and states to Pacino that there’s no afterlife but earth and worms, but then in succeeding novels other characters start having near-death experiences, like when Peter Vornado is suffering from cancer and even Pacino himself has an NDE when he half-drowns after the cruise liner he is on gets sunk. In Emergency Deep the entire crew of a French sub gets shot by terrorists and there’s a scene where they all meet up in Heaven and ask each other if it hurt when they passed away! In the same book, the departed wife of BK Dillinger, a new character, appears before him as a ghost and gives him advice. The idea of communication with ancestors and seeking their guidance is of course a quite accepted aspect of Shamanic culture and I wonder where DiMercurio has been getting his inspiration. Has he been hanging out with Native American philosophers? He does talk on the USS Devilfish website about hard times he’s gone through, being divorced, his friend dying. Often experiences like these open new avenues of thought in a person’s life.

Michael DiMercurio has not written anything for several years now and, based on his schedule up till now, he's overdue for a new novel. I'll be watching his space very keenly for when it comes out! I'm also interested in Kevin Warwick's career and what new ideas he comes out with. Here's a recent TV spot about Warwick, and unfortunately he's still unaware that he's being used to promote the Cyborg Agenda of the New World Order: http://morleyparents.blogspot.com/2008/08/kevin-warwick-encouraging-scared-mother.html As per usual, the perceived fear of crime is the stick that beats us. The mother of the girl in the vid, and the girl herself, ironically take the chip because “I just want to be free”. This way we will not need persuading; we will demand that these psychotronic weapons be used on us. The Illuminati want to do the opposite of the AI designers: turn entities that are naturally intelligent into mindless slaves like silicon computers.