介绍:
【Geek Time】-“极客”时间，和“科技宅”Brad一起聊神秘有趣的“黑科技”“高科技”“硬科技”。欢迎关注公号【璐璐的英文小酒馆】，可以查看更多精彩内容，查看英语全文稿哦~
Welcome back to Geek Time Advance. This is Brad. Welcome back, Lulu.
Hi, Brad.
So I thought we'd continue on with talking about some of the issues with ChatGPT.
I think in the previous episode, we focus mostly on how great it is, how different it is from some of the earlier chat bots. But there are some really obvious problems. One of the things is, have you read the news about some schools already starting to ban ChatGPT in school, in university, in tests?
One of the issues with ChatGPT is that when you give it a math problem, for example, it can give you the answer, but not only that, it can show you the work for the problem.
So all you have to do is just copy down what it tells you to do. And so you don't even have to go online and find if someone's answered the question before on a database somewhere. Because ChatGPT would work for you.
It's the perfect cheating machine.
Yeah. And you can also use it to write essays; and it goes online and it basically pulls information from all the stuff and then creates an essay for you.
Yeah, actually I watched a short video of a short interview with Chomsky, you know, 乔姆斯基, one of the most famous linguists and linguistic experts in the world.
He basically calls ChatGPT a form of high tech plagiarism.
You could say that.
That's what you're talking about, right？Writing essays, basically I can… as a student or as a researcher, I probably can just lay back enjoy whatever I enjoy and then let ChatGPT write my essay for me.
Exactly. Like people basically can write down information about their own experience. But rather than do that, you can use ChatGPT to go online and just write down other people's experiences for you as if it that your own.
But will you get found out？
Oh, yeah.
You know that some professors or some universities they have developed their programs to check whether you have plagiarized your work or not. 就是那种论文查重的.
What about essays written by AI, by ChatGPT?
They're creating AIs that can detect whether or not an AI took a picture or created a picture, I should say, or wrote an essay. There’re usually telltale signs whether or not something was written by an actual human or written by an AI.
Human beings. Why do we make life so difficult for ourselves？
And yet so easy.
I know. So we created AI technology like ChatGPT to write our essays for us, but then we have to create AI programs that will detect whether the essays are written by AI or not. It's a mouthful.
I'm also worried about the scams.
In deepfakes are a big issue nowadays especially as they get better and better. You could call someone and they could imitate your voice and look like you, and ChatGPT like software can be used to make deepfakes look more realistic.
I've seen some of the deepfake videos that is scary. So guys, if you don't really know what's that like, imagine that they can steal your face or they can create a human being that does not exist, like an expert that actually is not a real human expert, but then using AI technology, using deepfake, this person can look very realistic and they can use your voice as well, use your pattern of speech.
And now that they're using your voice and they're using your patterns of speech. They're coming for your jobs.
That's another concern, isn't it？People like, oh no, ChatGPT is gonna steal our jobs. What do you think？
Right now, ChatGPT is doing some things like editing code and can even write some python code.
So like it's getting kind of crazy what it can do. So it's probably not going to write high level code that people need for bigger programs. But definitely editing code is something that takes a lot of work. And it can help out with that.
It probably can replace certain jobs. I'm not sure if it's completely but like I said earlier, customer service, for example.
Yeah, customer service is a big thing. Data processing is something that's probably going
大家还在听
