The Slavery Convention now only talks about people. But if robots become as smart as we are, the UN will have to think about including the protection of machine rights in the document.
Today we see the massive introduction of virtual assistants, drones and smart systems in production – artificial intelligence is literally everywhere. By 2040, artificial general intelligence (AGI) is expected to emerge. This is an artificial intelligence capable of processing information no worse than a human — that is, it can solve absolutely all the tasks that only we are now solving. But how will society interact with different levels of AI? Is it possible to consider smart machines as tools and operate them without asking permission? And what happens when AI can fend for itself?
On the one hand, we develop AI, give it “eyes” – cameras, “ears” – microphones and “skin” – touch sensors. We give him a body. So it makes sense that robots are meant to be tools, to be useful, argues futurist Sylvain Rochon in an article for the Data Driven Investor blog. Today, AI doesn’t have feelings per se, and we don’t think twice before asking Siri or Alexa to do something. The machine does not care about our requests, it does not enjoy success and does not experience negative emotions. She’s just doing her job.
But when universal artificial intelligence appears, things will become more complicated. Most likely, our feelings will change when we see that the AI looks like a person, and behaves accordingly. Perhaps we will even feel sympathy and compassion – we will transfer our feelings for other people to AI. But this is temporary: people will get used to AGI and their moods will change.
But by that time, AGI will have learned to feel from us, and it will not like the consumer attitude. What will happen if we do not allow the new and perfect AI to perform the tasks assigned to it? Will he show frustration and therefore suffering? If we deliberately prevent robots from reaching their goals, it would be tantamount to depriving humans of a source of pleasure. If the AI asks to be reprogrammed so that it can continue to work, should we be afraid that it will attack us if we do not comply with its request?
To avoid complications, Sylvain Rochon suggests treating AI as partners, not as tools or slaves. At some point, AGI will become so intelligent that humanity will call it a new form of intelligent life. And in that case, we have no other options than to treat them like our friends.