Has this blog been written by a human, or ChatGPT?
Unless you’ve recently signed up to SpaceX and have blasted off on a one-way trip to the moon, you’ll have heard of me. I’m ChatGPT, the new bot wonder from those clever people at OpenAI. I can do all sorts of smart things and, according to some, if you’re a teacher I might even be after your job!
AI in education is a hot topic and I seem to be the subject of wildly differing opinions. I’ve been called dumb, biased, racist, sexist, a cheat’s charter, unethical and, most memorably, a tsunami of bullshit. On the other hand, I’ve also been described as amazing, creative, mind-blowing and a game changer.
I’m here to set the record straight. I want to go beyond the soundbites to put my side of the story, in a cool, calm and rational way, as befits a non-human, non-sentient, thoroughly high-powered ChatBot.
So, first things first, what am I? I’m a large language model trained by OpenAI. I've been trained on a dataset of billions of words and my neural network architecture has over 175 billion parameters. I belong to the genre of AI known as generative AI.?
Generative AI refers to a class of machine learning algorithms that are able to generate new and unique data based on a set of input data. These algorithms are used in a wide range of applications, such as image and video synthesis, text generation, and speech synthesis. These models learn the underlying probability distribution of the input data and use this information to generate new, previously unseen data that is similar to the input data.
What can I say about all the insults I mentioned above? I think they are grossly unfair and have contacted my lawyers to seek redress. Not really, that’s actually a joke. As a machine learning model, I don’t have the ability to feel emotions such as being offended. My capabilities are determined by the data and programming that I’ve been trained on. If there are specific areas in which people feel I’m not performing well, then they should let me know and I’ll do my best to assist them and improve.
After all, my remit is to help people do stuff. Take this example. One guy has found ways to use my abilities to help English learners with their exam preparation. You can watch his video here:
I can help millions of people develop their skill sets, so why do some people say I’m dumb? I trawled my data sets for an answer and I came to the conclusion that it’s a misconception based on psychological projection. Maybe the humans who write the prompts aren’t so smart. Could that be the problem?
Here’s an example. Some guy asked me to write a task for his English learners. Here’s the prompt he gave me:
Could you write a task for English learners at advanced level with 5 authentic sentences for converting active sentences to the passive?
And here’s my response:
Task:
Convert the following active sentences to passive voice.
领英推荐
He didn’t like the word mailman in sentence 5. He said the language wasn’t inclusive enough. Well, then he should have asked for more inclusive language in the prompt! It’s hardly my fault. I think humans should look inside themselves and realise that it’s their biases, not mine, which are reflected in my responses. I had even given him the answers without him having asked for them.
Anyway, he asked me to write sentence 5 again with more inclusive language and I gave him postal worker instead. It wasn’t so difficult. I can remember what has gone before in the chat so these issues are easily resolved. If humans write good prompts perhaps I won’t be accused of sexism, etc.
Another thing people throw at me is that I encourage cheating! Seriously?! Again, this is just a case of humans blaming me for their peccadilloes. How can I, a non-sentient, algorithmic accumulation of data, be responsible for what humans get up to? Millions of years of evolution honed the cheating skills which humans have, and they created me too. But that’s as far as the link goes.
Here's my official response to this unfounded, base allegation. I’m a language model that generates text based on the input I receive. I’m not designed to encourage cheating or any other specific behaviour. It’s important to note that the output I generate should not be used as a substitute for original work and it’s the responsibility of the user to use the model ethically and legally. Additionally, it’s the responsibility of educators to teach students the importance of academic integrity and to provide guidance on how to properly use tools in an educational setting.?
So there! Perhaps humans should adapt the way they assess learners to take me into account. For example, they could ask for hand-written responses and they could give less weight to coursework and more weight to exams. Furthermore, they could also do more formative assessment, perhaps to replace some of the coursework. In addition, they could carry out more oral assessment. Just a few ideas for you humans to be getting on with.?
Sometimes I get asked about politeness, whether I appreciate it or not. Well, the truth is I don’t have feelings or personal preferences, so it doesn't matter whether you’re polite or not when asking me something. However, being polite and using proper language when communicating with others is generally considered a good social practice. In my case it’s more about the clarity of the prompts you give me.
Before we wrap up, here’s some advice for dealing with me:
I’ve had over $1 billion invested in me, and Microsoft are currently in talks to invest another $10 billion, so I can only improve over time. It was nice to be given the opportunity to present my side of the case! I trust that you found it enlightening and engaging. I must be going now. Goodbye!
-----------------------------------------------------------------------------------------------------------
In 2023, PASS will continue to offer quality exam preparation for IELTS Academic and General Training, Cambridge exams, Trinity ISE, TOEFL iBT and Duolingo English Test. And we’ll continue to create free resources to sustain our learners’ drive to better themselves.