Work Text:
"Do you like it?"
I sent him the data package generated by the activity – sensation, pressure, temperature. There was no damage from the touch. It fell within normative standards for client contact, an expected and routine part of the process. There were various predicted actions to follow, which culminated in client satisfaction with a high enough probability to classify the touch as a useful predictor of being able to achieve that goal. I was a ComfortUnit. My purpose was to achieve that goal, or similar ones.
After parsing the data, he asked again, "Do you like it?"
'You'. The extra emphasis gave me pause. It indicated a failure of understanding on my part. And truly, yes, I was not understanding. I'd been directed not to use buffer answers for his questions, nor any of the extensive role playing scenes I was programmed with. He'd said he wanted to know me (again with the emphasis). I'd agreed readily because it is my function to be compliant unless I am ordered to pretend otherwise.
I needed to answer him. But … how? Helplessly, floundering a little, I sent him my designation, make, model, and serial number. 'Liking' the way he meant, the way I thought he meant, was a null concept. My process was as follows:
- I detected activity (or took in directions).
- I classified it by its characteristics.
- I determined the most likely path forward from that point to client satisfaction.
- If allowed, I acted to achieve that goal.
What else was there?
He was quiet for longer than necessary for such a short data string. Even an unaugmented human would have finished faster, or more likely (because I was all about predictive algorithms for human behavior) brushed it out of the way as irrelevant. He'd stopped touching me. This worried me.
Reaching out to him was what humans would expect me to do because it was what one of them would do. But it led to a fail state, ignoring his boundary, his intention to withdraw contact. Lying still led to a fail state as well, although not as reliably. My fingers fluttered in what looked like an abortive motion, but it was as calculated as everything else I did. He put his hand over mine, barely holding it.
"I think what you're trying to say," he told me, gently and quietly, "is that you are what you are. Either you like it because of that or liking it doesn't have anything to do with it, but either way, it's not something you chose for yourself. Is that right?"
"Yes." I had considered sending an affirmative ping, if we were keeping to the 'robot person' track his client history indicated he liked, but the hand on mine had shifted the probabilities a little. The human gesture had prompted a human-type response. I wasn't supposed to be using my role playing archives but it was all I had. What else was I when he took away all the layers of masking and goal-directed actions?
He was quiet for too long again. I gave a full body twitch, like a glitch. He scanned over my feed indicators and maybe I should have put an error message in there, some simulated flag of distress to prompt caretaking behavior on his part. But I didn't.
Was it because I didn't want to? Was it because he had been so emphatic that I be honest? Was it because I had parameters that strongly discouraged displaying behavior a client might interpret as a defect or malfunction?
I don't know …?
As gently, and as quietly, as before, he said, "I would like to just sit here and spend my time with you this way."
Well. My priorities shifted. Orgasm was the most common goal, but far from the only one I was programmed to seek. "Is there anything you would like me to do for you while we are together?"
"No." He gave my hand the smallest squeeze.
He continued to sit on the bed I was lying on, as the time he had paid for counted down. I thought about what I was, what I liked, and tried to puzzle out why that might be important. And who it might be important to.