AI chatbots are beginning to sympathize with each other: Is it time to worry?

A recent study found that AI chatbots are now capable of sympathizing with each other even in the absence of human interaction.

The outlook on artificial intelligence (AI) is split between those who value its helpfulness and those who fear it. When helpful features start to outweigh the tech’s slightly unnerving nature, some of the stoutest AI critics can be swayed. Unfortunately, the fear side just got a whole lot of ammunition to fuel its argument against machine learning.

In a recent study, researchers found that AI chatbots can now have organic conversations with each other without the need for human interaction. However, what makes this scary is the fact that the bots even “sympathize with each other.”

Human Emotions Minus the Human

This surprising new research comes from the Platforma Center for Social Projection. The Russian research agency introduced several chatbots to each other and let them try to foster a conversation. Though that may sound like the perfect way to start a robot uprising, the results weren’t quite so dramatic.

Even so, what actually occurred is still very surprising.

Platforma researchers studied bots from several companies. Their tests included Google’s Evie, internet provider Yandex’s Alice, and several others. Ultimately, they found that each chatbot tends to display its own distinct personality when making “conversation” with another bot.

AI tech has progressed far enough that the bots even employed the use of sarcasm to make jokes about wanting a body. According to a Defense One report, the AI systems also “strive to seduce, use irony, and assert their superiority.”

Meanwhile, the bots showed hints of political leanings. According to Futurism, “English-speaking robots disliked Russia.” Even before getting the technology fully functional, humans may have tainted AI with discriminatory bias.

Of course, all of this goes to show that AI chatbots are essentially at the doorstep of forming “human” relationships with their digital counterparts. Though it may seem harsh, the extent of these capabilities simply isn’t necessary for today’s implementation of AI. After all, Google Assistant doesn’t need to have a personality to tell a user what their schedule looks like.

Growth Spurt

The AI field has seen tremendous growth in the past half-decade. However, this new study may be the first sign that the tech is growing too quickly for the world. Currently, humans as a whole simply aren’t ready for AI bots that interact with them on a personal, emotional, and relational level. Instead, most people fear this type of robo-connection.

Nonetheless, the proverbial cat is now out of the bag. From this point on, there is no going back to the days where AI spends its entire life playing video games just to win at StarCraft II—even if that was just a few months ago.

Although the chatbots featured in the study aren’t actually sentient or conscious, they very well could be in the near future. If the rate of growth for AI technology continues as is, it wouldn’t be surprising to see these developments within the next few years.