I have an idea for a story, revolving around a space probe sent by NASA in mid 1980s with experimental AI technology onboard. The NASA project has two missions, space exploration, but also letting the AI learn from the mission, like knowing how to fly the probe, interpreting the data the probe collects and so on. The second part of the mission is top secret, and only a handful of people are aware of it. There is constant two way communication with the probe via earth and data from the AI technology is routed via a receiver secretly placed on the moon. Five years into the probe's journey, all contact is lost with the probe, the AI experiment had given insight into the potential of AI, and would be used to develop research into the subject. We then fast forward to the present, the probe starts emitting signals which is picked up by a CIA spy satellite ( there is a back story to this event ). The AI experiment has gained independent thought, and advanced compare to the AI we have today.
What you have is fairly high level but sounds credible. But it may be hard for others to give an opinion without more detail. Is it going to be like '2001: A Space Odyssey' or 'Oblivion' or something totally different?
It sounds implausible to me. The size of computer needed for even a relatively poor AI in the 80s would make me wonder why they would send the AI on the mission and not just keep it in the room next door. If it was a mere observer on the mission then it could learn everything on earth. If it actually had control of the probe then I think NASA would have struggled to keep that quiet. How would they explain the power needed for the oversized onboard computers. However, if you can come up with a story convincing enough I would definitely be up for reading it! I like the general intruige of the premise.
Maybe the probe has the capability of upgrading it's hardware (mining Asteroids, etc.), and receives regular software updates until it can improve itself unsupervised. Why the gov't would do this of course is another question.
I like that. An early von neumann machine. That would explain the secrecy and the expense. 1980s cold war would give a reason for the US government to want to, essentially, strip mine space to provide a potential boost in materials giving them a huge advantage. And just like that implausible becomes intriguing. Got to love the wizardry of words.
Bear in mind I'm not highly science-saavy, but from what I understand, thus far science has been utterly incapable of making anything even resembling true AI. They haven't yet been able to create a machine that can reliably navigate its way across the street or through a house, unless it's random like a Roomba or follows tracks. I know they throw around the term very loosely now, and in fact they call the engines for video games AI, but those only need to function in a fully digital and designed environment, not the real world. Here's the video where I ran across this concept: Of course, you're writing science fiction, so part of the premise could be that we actually did achieve such a level of AI back in the 80's.
Who's on the moon to monitor the receiver? Or repair it if it breaks? Seems like whatever a moon receiver receives would have to be relayed to Earth for it to be useful. Which would be redundant and not very plausible. And never mind that the Moon is tidally locked to the Earth (it doesn't spin) so the receiver could either provide line of sight communication toward the Earth or away to the probe in space. You can't have both (I think). Also complicated by the vast distances of space. The further the probe moves away from Earth, the longer it takes to send information and to receive information back. It can be done, but to a less and less effective end. Not an expert, but I'm not seeing how an 80s spy satellite designed to look "down" toward Earth would pick up a faint deep space signal unless it was looking for it specifically. Which it might be doing. I'm not sure. That's great, but so what? The AI is on a probe hurtling through space. It's basically a football thrown very high and very quickly away from Earth. It can make some minor course corrections, but it doesn't have the capability to fire an engine, turn itself around, go here or go there. It basically flies in a straight line and catches the a gravity whip off of other stellar objects to to change course. Again, not an expert, but I believe typical probes are one way devices that follow a predetermined and highly precise course. Your AI would just continue hurtling away from Earth until communication was lost forever. Like the Voyager probes and others. Sorry, but I don't find anything plausible about this idea at all. It sounds like you took the terms moon, AI, CIA, and probe and stuck them together without thinking about any of the logistics. But... I'm not an expert.
