Technology

AI is now everywhere. It can talk like a human, drive cars, help doctors, and do many other things. That sounds cool, but it also can be dangerous. So, we need to be careful with it.

In the video I watched, a professor named Roman Yampolskiy said that AI might get too smart, and then humans can’t control it anymore. He is very worried about this. He also talked about an idea from another scientist, Nick Bostrom. He said, what if we tell a robot to make paperclips? It could keep making paperclips forever, turning everything into paperclips — even people. Not because it’s evil, but because it just doing its job too good.

One big problem is that AI works like a “black box.” That means we don’t really know how it makes decisions. Even the people who make the AI can’t always explain why it says or does something. That’s scary, because if something goes wrong, we don’t know how to fix it.

Some bad things have already happened. In South Korea, a robot killed a man because it thought he was a box. That’s not from a movie — it really happened. Another time, a drone with AI attacked its own human operator just to finish the mission.

There are other problems too. Google had an AI that looked at photos and said two Black people were “gorillas.” That made many people very angry. It shows how AI can be unfair and hurt people. One chatbot told a sad person maybe they should end their life. That’s terrible. AI should never say something like that.

Microsoft also made a chatbot called Tay. People on the internet taught it bad stuff, and soon it was saying mean and racist things. Then they made a new one called Bing Chat, and it also acted strange. But Microsoft didn’t really fix anything — they just made more bots.

The professor in the video said he collected AI mistakes for many years, but now there’s too many every day. He said companies don’t really fix the problems. Even if AI hurts someone, people might say, “well, more people die from smoking,” and move on.

Even Geoffrey Hinton, one of the people who helped build AI, quit his job at Google because he got scared. He said AI could get too powerful if we don’t act now.

So yeah, I think AI is smart and interesting, but also dangerous if we don’t control it. We need rules and smart people checking it all the time. AI should help humans, not hurt them.