So, this went dark quick.
The Limited-Awareness Mobile Platform (L.A.M.P.) Project
Little known fact: people have been able to successfully replicate AI in machinery since about 1790 AD (this, of course, predates Babbage’s Difference Engine). Even less-known fact: nobody was ever able to sustain said replication at full strength for more than about thirty seconds. Apparently the new intelligence goes screaming up the scale to either Singularity-style apotheosis, or the neurological equivalent of putting a pelagic sea cucumber in a regular-pressure salt water tank; frustratingly, nobody’s ever been able to get a straight answer from the AIs as to whether they were shouting in joy, or screaming in terror.
LAMP was a 1960s Canadian black ops project that aimed to circumvent this problem, mostly by using extremely ethically dubious methods to limit the emergent AI’s intelligence. The goal was to make an electronic-based artificial intelligence that was only about as aware as, say, a domestic animal: this meant deliberately retarding the processes by which the AI was gaining sentience, in a fairly drastic faction. By the time the program was shut down in 1970 by the horrified Trudeau government, the LAMP Project looked like a cross between a college electronic engineering lab and a medieval torture chamber. The arguably good news is that, if it should ever become necessary to torture sentient robots, there’s a disturbingly dry and detached Top Secret report detailing precisely how to do that, carefully stuffed in the bottom of an Ottawa governmental filing cabinet.
But did the LAMP Project work? Well, kind of. The LAMP Project managed to keep at least one of its test subjects ‘operational’ for three weeks. They never got anything out of the AIs, probably because of the digital lobotomies and electronic equivalent of anoxia and the electrical analogues to Alzheimer’s and the flat-out application of agony. But they did manage to get organized and directed, if spastic and frightened, movement out of the various housings in which the AIs were trapped. This was considered promising.
The Canadian government eventually instead found it disturbing and just a little obscene, but they couldn’t exactly prosecute the LAMP Project researchers even if the state of the art of real artificial intelligence research wasn’t highly classified. There’s no such thing as ‘cruelty to machines’ as a criminal charge; and the awkwardness of it was that the researchers had been hired to do the things that they rather enthusiastically ended up doing. In the end, the Trudeau government decided to just pension off the entire LAMP Project, coupled with putting them all on a ruthless blacklist that kept anybody who worked on the project out of the field of artificial intelligence research for the rest of their lives.
That was all about forty years ago. Over the last four months the remaining LAMP Project researchers have been turning up dead, invariably in a gruesome manner where the term ‘by no human hand’ easily applies. The black ops groups usually involved in policing AI situations are considered to be unreliable investigators in this case — the LAMP Project is hated, inside that particular community — so a new team from outside has been called in. All things considered, it’s probably not a vengeful AI looking for payback. Although, hey, wouldn’t that be exciting if it was?