Artificial Intelligences Have Started To Attack Each Other In New Experiments

Not a dreamy situation.

Will, get out of there while you still can Image © Twentieth Century Fox. All rights reserved.

Can you imagine two robots fighting each other right in front of you? Well, don’t be too shocked, because it might happen sooner than you think.

Okay, perhaps not with robots per se, but it is something that can definitely happen with your computer or phone system.

DeepMind, the artificial intelligence company owned by Google, has conducted an experiment to analyse AI units and the way they react under certain circumstances.

For the first experiment DeepMind conducted, a game was created in which two AIs (one blue, the other red) had to collect apples. The AIs had the choice of shooting each other. As long as the number of apples was high, the AIs didn’t threaten each other, but the moment apples became scarce, both blue and red began shooting each other to obtain the apples.

For the second game, things got nicer. Two red dots acted as wolves that had to hunt their prey. The wolves could get bonus points if they stood close to each other, so the AIs decided to collaborate to get more points.

Creepy? The researchers explained that “the research may enable us to better understand and control the behaviour of complex multi-agent systems such as the economy, traffic, and environmental challenges.”

So yes, this research might have been used to understand human behaviour, but don’t be surprised if in a near future, your laptop and your phone start fighting and sending viruses to kill each other. Or even worse, they might form an alliance against you and start sending you subliminal messages while you’re using them. Trust no one.

Previous Post
Next Post