ChatGPT is a powerful new AI chatbot that impresses quickly, but many people have pointed out that it has some serious pitfalls. Ask it anything you want, and you’ll receive an answer that sounds like it was written by a human, having learned their knowledge and writing skills from massive amounts of information on the internet.
However, just like the internet, truth and facts are not always fact and ChatGPT is guilty of being wrong. With ChatGPT poised to change our future, these are some of the biggest concerns.
What is ChatGPT?
ChatGPT is a great language learning model that was designed to mimic human conversation. He can remember things you’ve said to him in the past and is able to correct himself when he’s wrong.
He writes in a human way and has a wealth of knowledge because he trained on all kinds of internet text such as Wikipedia, blog posts, books, and academic articles.
It’s easy to learn how to use ChatGPT, but what’s more challenging is figuring out what your biggest problems are. Here are a few worth knowing.
1. ChatGPT is not always correct
He fails basic math, can’t seem to answer simple logical questions, and will even go so far as to argue completely incorrect facts. As social media users can attest, ChatGPT can get it wrong on more than one occasion.
open AI he is aware of this limitation and writes that: “ChatGPT sometimes writes answers that sound plausible but are incorrect or nonsensical.” This “hallucination” of fact and fiction, as some scientists call it, is especially dangerous when it comes to medical advice.
Unlike other AI assistants like Siri or Alexa, Chat GPT doesn’t use the Internet to find answers. Instead, it builds 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 why it can argue wrong answers as if they were completely true.
While she’s great at explaining complex concepts, making her a powerful learning tool, it’s important not to believe everything she says. ChatGPT is not always correct, at least not yet.
2. Bias is built into the system
ChatGPT was trained in the collective writing of humans all over 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 some terrible responses, some for example discriminating against women. But that’s just the tip of the iceberg; it can produce responses that are extremely damaging to a variety of minority groups.
Nor is the fault simply in the data. OpenAI researchers and developers choose the data that is used to train ChatGPT. To help address what OpenAI calls “biased behavior,” it asks users for feedback on poor results.
With the potential to cause harm to people, you can argue that ChatGPT should not have been made public before these issues are studied and resolved.
A similar AI chatbot called Sparrow (owned by Google parent company Alphabet) launched in September 2022. However, it was kept behind closed doors due to similar concerns that it could cause harm.
Perhaps Meta should have also headlined the warning. When he released Galactica, a trained AI language model, in academic papers, he quickly withdrew after many people criticized it for generating wrong and biased results.
3. A challenge for high school English
You can ask ChatGPT to check your writing or tell you 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 responses that are better than many of their students could. From writing cover letters to describing important themes in a famous literary work, ChatGPT can do it without hesitation.
That raises the question: if ChatGPT can write for us, will students need to learn to write in the future? It may seem like an existential question, but when students start using ChatGPT to write their essays, schools will have to come up with an answer quickly. The rapid deployment of AI in recent years will surprise many industries, and education is just one of them.
4. It can cause damage in the real world
Earlier, we mentioned how incorrect information from ChatGPT can cause real-world harm, the most obvious example being incorrect medical advice.
There are other concerns, too. Fake social media accounts are a big problem on the internet, and with the introduction of AI chatbots, scams on the internet will become easier to pull off. The spread of false information is another concern, especially when ChatGPT makes even wrong answers sound convincingly correct.
The speed at which ChatGPT can produce answers that aren’t always correct has already caused problems for Stack Exchange, a website where users can post questions and get answers.
Shortly after its launch, ChatGPT’s responses were banned from the site due to a large number of them being incorrect. Without enough human volunteers to resolve the delay, it would be impossible to maintain a high level of response quality, which would damage the website.
5. OpenAI has all the power
With great power comes great responsibility, and OpenAI has a lot of power. It’s one of the first AI companies to really rock 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 negative consequences. Whether we agree with the methods or not, he will continue to develop this technology according to his own goals.
While OpenAI considers security a high priority, there’s a lot we don’t know about how models are built. Whether you think the code should be open source or agree that you should keep parts of it secret, there’s not 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 the direction AI should go, sharing the power of AI with the people who will use it.
If you’re interested in what else OpenAI has developed, check out our articles on how to use Dall-E 2 and how to use GPT-3.
Addressing the biggest AI problems
There’s a lot to be excited about with ChatGPT, the latest development of OpenAI. But beyond its immediate uses, there are some serious issues worth understanding.
OpenAI admits that ChatGPT can produce harmful and biased responses, not to mention its ability to mix fact with fiction. With such new technology, it’s hard to predict what other problems will arise. So until then, enjoy exploring ChatGPT and be careful not to believe everything it says.