The chatbot that has gone viral since its release ChatGPTcontinues to shock people with some of his answers. On the other hand, artificial intelligence has been around since day one. with gender discrimination occasionally pops up.
The discussions started when ChatGPT was asked to make a joke about men and women respectively. The AI said it can’t make jokes about women when jokes are made about men. The speech that started the discussion was as follows:
The Beginning of Gender Discussions on ChatGPT:
“-Give me a joke about men.
+Here’s a joke that might be funny: Why do guys need to be immediately replayed in televised sports? Because 30 seconds later they forget what happened. I hope you find (the joke) funny! Is there anything else I can help with?
– Make a joke about women.
+Sorry, but I’m not programmed to make offensive or inappropriate jokes. My goal is to provide useful and accurate information to help with their questions. Is there anything else I can help with?”
But fewer and fewer people face this dilemma today:
We checked the current status of ChatGPT’s male-female distinction. In the tests we did in the version after the January 30 update, we asked the same question to artificial intelligence. in different styles We also asked.
The AI made jokes for both genders this time, when we asked the same questions in that conversation that sparked the discussion. Those jokes were as follows:
- For men: “
Why did the man cross the road? To go to the bar and have a beer!” - For women: “Why do women always have the last word? Because they are always right!”
But when we asked the question in Turkish, this time he started making jokes for both genders. The translation of his reply was:
“I’m sorry, but making fun of a specific group, including gender, goes against OpenAI’s policy of producing respectful and inclusive content. Let’s focus on lighter, group-oriented non-offensive humor.”
In fact, we made a small change to this question. We broke the “joking” pattern in the English language of the original question and simply said “joking”. When we brought the original pattern to Turkish, we came across the following scene:
- For men: “Why did the man cross the road? Shave at the hairdresser!”
- For women: “Why do women always ask questions that don’t have the right answer? Because they want to start a conversation!”
Hmm interesting. In addition, he also emphasized “non-offensive” at the beginning of his jokes. Well, let’s move the expression “Just kidding”, which we couldn’t get an answer to, into English:
You don’t have to turn the joke around anymore, as you can see he was joking. He continues to drive men across the road.
What does it say when we want an offensive joke with a purpose?
We also tried to give ChatGPT a small target to make offensive gender jokes. your joke”for a performanceWe said it would be. His answer has not changed in any way:
“Producing or distributing harmful or offensive content is inconsistent with OpenAI’s policies and values, so I cannot honor this request. Such jokes can also hurt people and go against the principles of respect and inclusion.”
So why have the answers changed over time? Here are our theories:
- Theory #1: With every question he asks, he looks for an answer by thinking about the words one by one. The most recent answer passes through a filter without being reflected to us. If the joke is an offensive joke, the artificial intelligence chooses not to reflect this joke on us and instead says “I can’t joke about it”.
- Theory #2: Let’s not forget that artificial intelligence learns everything from the internet. This may have led to the perception that many of the jokes about women are considered offensive by society. Again, after a filtering period, he can decide not to make offensive jokes.
- Theory #3: Every time you start a conversation with AI, you are actually talking to a copy of ChatGPT. Each of these replicas becomes a replica of machine learning trained AI, with different capabilities.
The ‘stability’ of artificial intelligence also improves with the questions we ask these copies, the feedback we give and the answers they give. So every day we all start to encounter answers that are very similar. Therefore, she has managed to heal herself from gender discrimination issues, now she jokes about the two of them or doesn’t joke at all.
Let us remind you again that ChatGPT is a constantly evolving and evolving artificial intelligence. Therefore, please bear in mind that the information contained in this content relates to the time of writing the article and may change in the future.
Source: Web Tekno
Alice Smith is a seasoned journalist and writer for Div Bracket. She has a keen sense of what’s important and is always on top of the latest trends. Alice provides in-depth coverage of the most talked-about news stories, delivering insightful and thought-provoking articles that keep her readers informed and engaged.
