The Reverse Test

He was going to be late for his interview, so he stormed past the security and straight into the elevator. 37th floor.

When he arrived he could barely make it out of the elevator before an angry man almost ran in. Probably the previous candidate, and probably the answer was “no”.

Recently interviews have become a “yes” or “no” test. It was a weird phenomenon that started getting mainstream after a couple of robots posing as people got jobs in high profile companies and then murdered many of their coworkers. Robots were getting infected with viruses and quarantined or exterminated regularly, but many were still walking around in their humanly outfits.

And you couldn’t tell if the person in front of you is a robot or a man, and you weren’t allow to perform any “probing” — the government banned it after a few companies were stripping their candidates and even punching their arms and legs to see if there’s blood. Robots had learned to get passed these anyway, including electromagnetic field scans, with very realistic human bodies and electromagnetic shields, so it was both rude and useless to test them physically.

Interviewers had one important task — to find whether the candidate is a human.

“Hello, my name is Jane, how are you?”

“Very good, thank you. And sorry for being late.”

“No problem. Take a seat. You are applying for … a project manager, right?”

“Yes”

“OK. I will be asking you a few questions. Then we will know whether you can proceed to the skill interview. I’m sure you know the procedure, so I’ll just get started.”

The questions were very diverse and included variations of various ethical dilemmas, summarizing a short story, drawing conclusions in hypothetical situations, explaining a complex concept with simpler analogies.

He was well prepared as he knew that interviewers test the cognitive skills of the candidate — whether they could do analysis and synthesis, whether they could cope with abstractions and analogies. They also presented ethical and moral thought experiments to verify the inherent human uncertainty in purely utilitarian approaches.

An hour later he got a “yes”, followed by half an hour of technical questions, and he was then offered to job. While waiting for the elevator he asked:

“ What is your rejection rate? On the reverse test, I mean.”

“Around 90%”

He was very surprised. The official data was that robots were no more than 10% of the population.

“Are there that many robots?”, he asked in disbelief, “Or is the test flawed?”.

“No, robots are probably a lot less. But we can’t be sure. You see, many people don’t seem to have these cognitive skills. You cannot easily tell whether they are a very good AI or a pretty sloppy human. They talk in cliches, they can’t manage to grasp abstractions, their analogies are as if taken out of a very simple word mapping, rather than a product of conceptual understanding. And their ethical stances are quite simplistic, often boiling down to a few extremes.

But the test isn’t flawed. It certainly rules out the robots and has a bonus of ruling out the candidates that would most likely not do well anyway”

The elevator doors opened, he said goodbye to the lady and pressed the ground floor button. Thirty seven floors were enough for him to realize one scary thing — it appeared that the vast majority of humanity had reached the point where it is no better than artificial intelligence and had lost its human cognitive abilities.

The reverse test, which was short for “Reverse Turing test”, now practically served both as a safety measure and an IQ test.

The thought haunted the rest of his day. He was trying to figure out how things had ended up that way. Maybe it was a conspiracy, or a consequence of the modern lifestyle? Maybe it was that technology was developed before people’s brains could catch up? Or maybe it was always like that?