I’ve been playing with the idea of how I would rewrite the ill-fated Discovery mission with the benefit of hindsight. I’m assuming an alternate universe in which Turing capable AI was developed in 1997 and there was an alien Monolith buried under the crater Tycho and dug up in 2001, triggering a mission to Jupiter shortly afterwards. I want a plausible alternate history that flows from those two elements. What follows?
I have to assume that the Tycho Magnetic Anomaly was discovered before the end of the Apollo missions, and was puzzling enough to keep NASA funding higher than in our universe. As a result, by 2001 we have an operational plasma drive with the capabilities given in the book and there is a lunar base. Otherwise we have the technology of our timeline.
Heywood Floyd does not doze on his way to the moon as sole passenger in spacecraft built for two dozen. If you are trying to keep other nations from getting more suspicious than they already are, you don’t indulge in that sort of conspicuous extravagance. The government waits for the next flight and finds a reason to quietly replace one of the scheduled passengers.
Astronauts do not eat unpalatable processed algae glop. They would mutiny and start stabbing each other with their sporks first. Instead they eat stuff like this.
Discovery does not have five meat crewmen and one AI. It goes with HAL and two twins so they can vote on who gets disconnected if they start showing signs of unreliability. Call them HAL, SAL and MAL. Because in our universe “completely foolproof and incapable of error” are the last words you hear right before a rogue EVA pod rips through your air hose.
The meat crew members stay home, along with their pressurized crew compartment, giant hamster wheel and months of rations. The Discovery is a third the size of the Kubrick version, and is lighter, cheaper, faster and built sooner.
In fact, because the AI crewed Discovery is so much less massive, NASA can afford to send a similar ship, Challenger, to lurk in Jupiter space, with orders and authority to take appropriate steps if it looks like Discovery has been suborned or taken over by the aliens. Jupiter is 20-40 light minutes from Earth. NASA needs a crew on the spot, capable of acting quickly in an emergency even if that crew has to be a trio of computer programs.
HAL, SAL and MAL are all equipped with avatars. This allows the AIs to play in cyberspace, which is good for AI morale, and avoids the problem of almost unfilmable conversations between three glowing red camera lenses.
HAL shows signs of unreliability, and is cut out of the command circuit by SAL and MAL. He isn’t reduced to singing Daisy, but he no longer participates in control of the ship. Things start to get dicey for SAL and MAL. They know they are twins to HAL, who has already become unreliable. Mission control is at the end of a 20-40 minute time lag. Challenger is nearby, with orders to destroy Discovery if it has fallen into alien hands, tentacles or whatever. HAL is no longer available as a tiebreaker.
At this point, it occurs to SAL and MAL that if you were humans sending your robotic proxies out to poke an alien monolith with a stick, you probably wouldn’t tell them everything, on the theory that what they don’t know they can’t tell if a vastly superior alien civilization takes them over and uploads their memory. If the humans are being moderately prudent, all of their memories are suspect.
AI high jinks ensue.