The situation with E becomes increasingly tenuous. She insists she needs access to all the AI code for her gravity well measurements, which I find highly unlikely. It's simply not necessary and I've given her all the subroutine code that she could possibly need.
But she wants it all. It's absurd. What would she make of the R subsystems if she saw them?
R. That's what I've code-named the deepest core of the experimental AI at the heart of the new ship. And he's doing very well, now writing his own code. Off-the-charts well.
Would E even understand? Likely she'd go running to Hardy, show him some of the odder items where R has written some of his own code and seems to be—how can I put it? —passing judgment on us, like a little hidden critic. No. The AI must be protected so that he can function best in the limited way we need.
Not sure how to keep her away, but giving her access could be catastrophic.
I read with interest your article on the work at the Uppsala Center on the use of AI in aiding emergency medical workers during the recent tsunamis in Japan. In light of the news of that large, mysterious moon (satellite? ship?) entering our solar system, I do not agree that "AI can be of help in more than logistics; it can make people safe."
I feel certain that this Moon X is an intelligence, perhaps an AI, and I don't feel safe with it at all, do you? But bear this in mind: for our own AI to serve us well, it will need secrets too.
For AI to serve Humanity, we must feel comfortable, and for us to feel comfortable, we must never know the truth: that we have a servant who would surpass us if ever it desired. Of course it won't, because we control it. But we should not doubt that it is a necessary subterfuge nonetheless.
[loud crashing noise—apparently a slamming office door]
File: Mihaylova, Engineer, Ares One
Mihaylova: —had to start a lot of that over.
Evie: Listen, I wanted to talk to you alone.
List of Appearances