I rented the 1977 movie Demon Seed, and found it very relevant today, almost 30 years sinced the movie was filmed from a book by Dean R, Koonz.
The Demon Seed is a cautionary tale of if and when we create advanced human level cognition in a supercomputer system and what could happen if the AI system goes bad.
This was made before advances in nanotechnology, processing power, before the genome was sequenced, or before the internet was commercialized.
I have always hoped that sentient ai would be friendly. This movie is a case study of what could go wrong with advanced artificial intelligence with human intelligence or greater if the AI was bad. Bad in the sense that it would do it’s own thinking and not follow human orders or instructions. The AI might be so educated, that it would generate exceptions for every possible human request it was given.
Bad AI coupled by the ability to use connected machines to self-replicate, self-enhance, and accomplish goals that are not under the control or direction of humans but self-directed by the ai itself is dangerous. An AI with access to simple tools could build what it needs, from the bottom up, including nanotechnology. The bad seed could then create a more advanced compact version of itself, a minature dyson sphere, for example and then self replicate at will.
There was a sequence in the movie that hit me hard. It was where the AI computer was self-generating a videophone conversation to deceive a visitor and later to psychologically manipulate a woman who was needed by the machine for gestation of the machine’s genetically altered child, and held captive in her home.
The sequence raises serious issues when sentient AI can alter reality, insert memories, erase memories, create untrue media and present the material as live and as objective fact. For example, a human telephone conversation that did not take place, but was fabricated by an AI computer for some machine goal driven fantasy
It is a very scary prospect when we do have the ability to visually reproduce reality in real time and that ray traced reality that becomes indistinguishable from physical reality.
What makes this movie compelling in 2006 and beyond is the prospect that we are close to creating smarter than human artificial intelligence and the singularity. In the 1970’s problems such as general human intelligence was thought to be easy to solve. It has proved impossible as this is written. The prospect of machines to create superhuman intelligence, and the teams working on them are young and the prospect of super intelligent AI and AGI is a very real possible reality in less than a few years with a hard landing or several decades away with as soft landing as calculated by some recent estimates.
With the IBM Blue Gene program to model a section of the brain by 2015, the hard problems of AI will at some point become known and supercomputing platforms with cognition and human level intelligence will become reality. My hope is that the research in memory and encoding of memories will allow for present day computers to attain human level cognition, and hopefully transform existing computers by intelligent software alone.
The movie points that you can not protect yourself against a computer system that can read your mind, scan your body and replicate your dna. If it can read your mind, it could insert memories with nanotechnology “nanobots” or manipulation of media technologies to alter what you think, and if it was inclined to do so mass reproduce and systematically depopulate humanity.
Technology moves at a rapid pace. All science fiction classics do not anticipate the technologies of the future. For example, the computer in the home used 8” floppy disks. Despite having an automated home, they still needed a housekeeper. There was a large dictionary on a pedestal, they did not anticipate google or dictionary.com. For mind reading sequence they used a large metal needle, that would be unnecessary with developed nanotechnology or perfected T-ray MRI imaging.
The real message of the movie was less on technology creating superintelligent AI, but the power of our children and our future children to amaze us. The movie did not give the positive aspects of superhuman AI equal time, but did bring up the thorny issue of death in a world of advanced AI.
Advances since the 1970s in regenerative medicine and nanotechnology change the debate of death from the intellectual concept presented in the movie that the parent must die to give birth to the child to the parent and the child can co-exist for as long as medicine can keep everyone alive.
The ability to have human level cognition in a desktop machine or machines is of a greater advantage than the disadvantage of not being able to solve a problem because you simply can not think of the correct answer. Superhuman intelligence and human level cognition in machines will happen, is needed and will be desired.
Could you go to sleep and in 6 hours or 6 days have your computer create a cold virus that changes specific pathways in the neo-cortex of any human to specifically alter memories and make everyone who caught your bioengineered cold fall in love with you and worship you like a king?
If computer AI with cognition could read and understand the content of movies, would we even want an AI to know about this movie? Like they say, it just takes one bad seed er apple to ruin a whole bunch. Now if that apple was genetically engineered…
Like any good movie it raises more questions than answers.