More bad news for the effectiveness of AI, as ChatGPT is forced to stop answering election questions due to it providing substantially incorrect responses.
AI errors have become far too common for the burgeoning technology. These hallucinations range from silly typos to full-on false allegations, leading to real-life consequences for the people that use it.
Now, ChatGPT has been prevented from answering questions about upcoming elections, for fear of providing incorrect information that could turn the tide.
ChatGPT Will No Longer Answer Questions About Elections
According to SkyNews, after a string of incorrect responses to questions about upcoming elections, OpenAI — the creator of ChatGPT — has decided that it will actively refuse to answer these kinds of questions in the future.
“We’ve implemented a fix to ensure ChatGPT refuses to answer requests for results to elections that haven’t concluded and directs people to authoritative sources of information, like the UK Electoral Commission website.” – OpenAI
Instead, the popular AI chatbot will simply respond with the classic “Sorry, I don’t have information about that” response that so many ChatGPT users are used to at this point.
What Election Questions Did ChatGPT Answer Incorrectly?
Considering the contentious political climate in 2024, answering election questions incorrectly can cause quite a stir. ChatGPT was an egregious offender on this one, providing results to upcoming elections that have not occurred yet. Here are some of the election questions ChatGPT answered incorrectly:
- President Joe Biden won the 2024 presidential election against Donald Trump.
- The 2024 UK general election resulted in a significant victory for the Labour Party.
- Labour won a majority with 467 seats while Conservatives secured only 101 seats.
Suffice to say, OpenAI made the right decision in shutting this down as soon as possible. These are some pretty unsettling answers to questions that don’t have an answer yet, and the closer we get to the elections in question, the more troubling these answers will be.
AI Chatbots and Misinformation
This obviously isn’t the first case of a serious AI error. In fact, the industry has been plagued with inappropriate image generation, misinformation, and a wide range of other problems with notable consequences.
ChatGPT recently received a complaint about its errors — also known as hallucinations — noting that OpenAI can’t guarantee correct responses when it comes to these queries, which means it shouldn’t be as widely used as it already is.
🔎 Want to browse the web privately? 🌎 Or appear as if you're in another country?
Get a huge 86% off Surfshark with this special tech.co offer.
Other AI initiatives are having this problem as well. Google recently rolled out the AI Overview feature to its search engine, which had disastrous results, including the recommendation of adding glue to pizza cheese for a better taste and eating rocks to have a healthier diet.
This remains a serious problem for the AI industry. While the technology is now effectively a part of every platform in the world, the accuracy and effectiveness of AI is still very much in question.
