You have almost certainly heard of ChatGPT and similar programs. These AI chat bots take a prompt from the user and use available data on the internet to produce responses. There are a lot of misconceptions around these bots, including how they work and what they mean for society. So, I’m here to help debunk them and tell you the real answers.
1.“Chat bots make their answers out of nothing.”
Well, no. That’s not true. These bots do not produce sentences from nothing. Instead, every response and piece of dialogue is a mashup of different sentences found on the internet and stored data cobbled together into a legible response. They do this by taking the input (the sentence you put in) and running it through every single other piece of data they have in order to produce an output (the sentence it sends back).
2. “Chat bots are becoming intelligent.”
Also no. This one is honestly just a misunderstanding of chat bots becoming more advanced. The “learning systems” that AI undergo are not as simple as just telling them facts. One example is a kind of deep learning (essentially, telling an AI whether it’s right or wrong when identifying an image) where several bots are made at random, then tested on, say, recognizing characters in an image. They are scored, the best performing bots are kept, and the rest are discarded. Then, more bots are made based on the best performers. Rinse and repeat.
3. “AI are becoming alive, like Hal 9000 or The Terminator.”
No. They aren’t. That’s not how AI works. AI, as we understand it, cannot live like Hal 9000 or make fully independent choices. As we currently understand AI, to become alive, it would have to be able to create data without some basis, which currently just isn’t possible. If that were to somehow happen, it would be decades in the future, at least.
4. “This is gonna steal our jobs!”
Eh, yes and no. It is entirely possible that fields like TV writing will be dominated by AI in the not-so-distant future, which is awful. However, basically no one else is in trouble, because AI are super specialized, and importantly, limited to only what a human can physically build. Your average factory worker will not have their job replaced within the next two decades, most likely. People like security workers and those in the military are completely safe. While this isn’t great for creative workers (in fact it’s terrible for us), it’s not the “oh-my-God-AI-job-apocalypse” that some people think it might be.
5. “AI becoming more advanced can only be a bad thing.”
This is a debatable sentence, because once again, only kind of. While AI becoming more advanced is bound to take its toll on creative industries like movies, it’s actually super useful for other things. Some examples include organization, inventory planning, optimization of stock in stores and websites. This also includes everyone’s favorite – personalized advertising, which is actually a decently old phenomenon (and is undeniably effective). But some more important examples include scanning financial transactions for fraud, cross-comparing symptoms and autonomous vehicles. And while these things aren’t perfect and do make mistakes, there are absolutely benefits to AI becoming more advanced. In school settings, there aren’t many uses that are directly relevant (aside from the popular use of AI to write essays, which I don’t recommend).
