A.I. - Artificial Ineptness

Hello test subjects, Cave Johnson here.

I find it hard to believe that it’s been a week since I last wrote for Felix – time sure does fly when you’re making the world a better place. Well actually, last week did go a bit faster than usual – I found my engineers playing with time again. You would have thought that they had learned their lesson the last time that they fiddled with time and ended up causing a rift in fabric of the space-time continuum, causing the [censored]. Again. That’s why point 83b of the Aperture Science Evacuation Guidelines now includes dinosaurs.

There was an upside though – we think our pest problem has subsided due to the roaming raptors, and those that we’ve caught are surprisingly good in the portal testing grounds. Time should be back to normal, but if you experience any time-related phenomena, such as déjà vu, then let your college test associates know. Bringing back an extinct species, debunking all previous theories about relativity and opening a new, niche pet-market? That’s us, and it could be you – summer placements opening soon!

What is your stance on artificial intelligence, and what dangers do you think it possesses?

Now, for those who missed the seminars, training sessions, RCSU Science Challenge and newspaper adverts, I’ve been asked by Felix to answer any questions you have about what happens here at Aperture Science. I’m also here to impart my wisdom and good-looks upon the latest generation of scientists and engineers. Not mathematicians – we don’t need you. This week, Martin asks, “What is your stance on artificial intelligence, and what dangers do you think it possesses?” Thanks for contributing to science, Martin – a voucher is in the post.

(Legal note: Aperture Science Redemption Vouchers can only be used to redeem an Aperture Science Redemption Voucher.)

We started working on a simple artificial intelligence a couple of years ago. The goal was to make the perfect scientist – one that didn’t have to be paid, wouldn’t complain, didn’t take breaks and would never stop testing. The bean counters started panicking when the project was first mentioned, due to the price of an actual computer being enough to bankrupt this company many times over. Instead, I got the boys in the lab to find a way to connect multiple smaller gizmos together to make a much more powerful computer than we could buy. That week, we took delivery of a couple of thousand calculators, put them in one of our brand new enrichment spheres, and linked them all up.

Artificial Stupidity – by learning to avoid more and more stupid things it effectively became more intelligent

Hesitation is just wasted time in the pursuit of science, and so we jumped straight in to making the most advanced artificial intelligence possible. Well, it turns out that it’s a lot easier said than done. The boys in the lab got as far as teaching the calculator to feel love, but the A.I. didn’t want to learn anything else. Their final reports stated that they tried to upset the A.I. but all that happened was the emotion altered from love to loneliness. The military would have paid top bucks for an intelligence that can only feel anger, but an A.I. that has abandonment issues is of no use to anyone. Most of the lab boys ended up building an emotional attachment to the experiment, and when it came to disassembly many girly tears were shed. Those who refused to participate in the clean-up were fired. From a cannon. Some even reached a high-point in their career – about 62 miles above sea level. The boffins say that’s technically outer space, so congratulations science-nauts! We’ll send your remains to your families.

The remaining boys in the lab gave up and said the idea of artificial intelligence was fundamentally flawed. By trying to create an intelligent being, we were limiting how much it could grow. Instead, we decided to work from the opposite direction and create artificial stupidity. We had the system evaluate its choices and it started to develop – not by being intelligent, but by not being stupid; by learning to avoid more and more stupid things it it effectively became more intelligent. The project was shelved last year after something went wrong and the artificial stupidity started making only the most stupid choices available. At least it was more intelligent than Dr Wheatley (the ex-head of the project) who defected to Black Mesa. Science finds ways to exact revenge on traitors. Incomplete formulae, decreased budgets and forced religion await all who scorn Science, when their time is up.

So, from what we’ve seen so far, artificial intelligence is far beyond our current capabilities, but don’t be afraid, concerned or dismayed test subjects. When Aperture Science makes the breakthroughs needed, we will try to make sure it doesn’t want to kill all humans. Well, we’ll think about it, at least.

As previously, any correspondence (questions, resumes or money) can be sent to felix.games@imperial.ac.uk with the title “Ask Cave”, and they’ll forward it on to Caroline.

A couple of testing raptors escaped with their Aperture Science Handheld Portal Devices, and have yet to be caught, so stay alert.

Cave out.