5 Big Problems With OpenAI’s ChatGPTReading Time: 5 minutes
OpenAI’s new chatbot has garnered attention for its impressive answers, but how much of it is believable? Let’s explore the dark side of ChatGPT.
ChatGPT is a powerful new AI chatbot that is quick to impress, yet plenty of people have pointed out that it has some serious pitfalls. Ask it anything you like, and you will receive an answer that sounds like it was written by a human, having learned its knowledge and writing skills from mass amounts of information across the internet.
Just like the internet, however, truth and facts are not always a given and ChatGPT is guilty of getting it wrong. With ChatGPT set to change our future, here are some of the biggest concerns.
What Is ChatGPT?
ChatGPT is a large language learning model that was designed to imitate human conversation. It can remember things you have said to it in the past and is capable of correcting itself when wrong.
It writes in a human-like way and has a wealth of knowledge because it was trained on all sorts of text from the internet, such as Wikipedia, blog posts, books, and academic articles.
It’s easy to learn how to use ChatGPT, but what is more challenging is finding out what its biggest problems are. Here are some that are worth knowing about.
1. ChatGPT Isn’t Always Right
It fails at basic math, can’t seem to answer simple logic questions, and will even go as far as to argue completely incorrect facts. As social media users can attest, ChatGPT can get it wrong on more than one occasion.
OpenAI knows about this limitation, writing that: ‘ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers.’ This ‘hallucination’ of fact and fiction, as some scientists call it, is especially dangerous when it comes to something like medical advice.
Unlike other AI assistants like Siri or Alexa, Chat GPT doesn’t use the internet to locate answers. Instead, it constructs a sentence word by word, selecting the most likely ‘token’ that should come next, based on its training.
In other words, ChatGPT arrives at an answer by making a series of guesses, which is part of the reason it can argue wrong answers as if they were completely true.
While it’s great at explaining complex concepts, making it a powerful tool for learning, it’s important not to believe everything it says. ChatGPT isn’t always correct—at least, not yet.
2. Bias Is Baked Into the System
ChatGPT was trained on the collective writing of humans across the world, past and present. This means that the same biases that exist in the data, can also appear in the model.
In fact, users have shown how ChatGPT can give produce some terrible answers, some, for example, that discriminate against women. But that’s just the tip of the iceberg; it can produce answers that are extremely harmful to a range of minority groups.
The blame doesn’t simply lie in the data either. OpenAI researchers and developers choose the data that is used to train ChatGPT. To help address what OpenAI calls ‘biased behavior’, it is asking users to give feedback on bad outputs.
With the potential to cause harm to people, you can argue that ChatGPT shouldn’t have been released to the public before these problems are studied and resolved.
A similar AI chatbot called Sparrow (owned by Google’s parent company, Alphabet) was released in September 2022. However, it was kept behind closed doors because of similar concerns that it could cause harm.
Perhaps Meta should have headed the warning too. When it released Galactica, an AI language model trained on academic papers, it was rapidly recalled after many people criticized it for outputting wrong and biased results.
3. A Challenge to High School English
You can ask ChatGPT to proofread your writing or point out how to improve a paragraph. Alternatively, you can remove yourself from the equation entirely and ask ChatGPT to write something for you.
Teachers have experimented with feeding English assignments to ChatGPT and have received answers that are better than what many of their students could do. From writing cover letters to describing major themes in a famous work of literature, ChatGPT can do it without hesitating.
That begs the question: if ChatGPT can write for us, will students need to learn to write in the future? It might seem like an existential question, but when students start using ChatGPT to help write their essays, schools will have to think of an answer fast. The rapid deployment of AI in recent years is set to shock many industries, and education is just one of them.
4. It Can Cause Real-World Harm
Earlier, we mentioned how incorrect information by ChatGPT can cause real-world harm, with the most obvious example being wrong medical advice.
There are other concerns too. Fake social media accounts pose a huge problem on the internet and with the introduction of AI chatbots, internet scams would become easier to carry out. The spread of fake information is another concern, especially when ChatGPT makes even wrong answers sound convincingly right.
The rate at which ChatGPT can produce answers which aren’t always correct has already caused problems for Stack Exchange, a website where users can post questions and get answers.
Soon after its release, answers by ChatGPT were banned from the site due to a large number of them being wrong. Without enough human volunteers to sort through the backlog, it would be impossible to maintain a high level of quality answers, causing damage to the website.
5. OpenAI Has All the Power
With great power comes great responsibility, and OpenAI holds a lot of power. It’s one of the first AI companies to truly shake up the world with not one, but multiple AI models, including Dall-E 2, GPT-3, and now, ChatGPT.
OpenAI chooses what data is used to train ChatGPT and how it deals with the negative consequences. Whether we agree with the methods or not, it will continue developing this technology according to its own goals.
While OpenAI considers safety to be a high priority, there is a lot that we don’t know about how the models are created. Whether you think that the code should be made open source, or agree that it should keep parts of it a secret, there isn’t much we can do about it.
At the end of the day, all we can do is trust that OpenAI will research, develop, and use ChatGPT responsibly. Alternatively, we can advocate for more people to have a say in which direction AI should head, sharing the power of AI with the people who will use it.
Tackling AI’s Biggest Problems
There is a lot to be excited about with ChatGPT, OpenAI’s latest development. But beyond its immediate uses, there are some serious problems that are worth understanding.
OpenAI admits that ChatGPT can produce harmful and biased answers, not to mention its ability to mix fact with fiction. With such a new technology, it’s difficult to predict what other problems will arise. So until then, enjoy exploring ChatGPT and be careful not to believe everything it says.