Artificial Intelligences (AIs) created based on advances in technological knowledge and language models are now widely used by the public, including in commercial and business sectors.
Their functions enable a multitude of useful actions that save people a lot of effort. However, the information provided by robots is not always completely accurate, which can cause a series of problems.
ChatGPT, OpenAI’s artificial intelligence, is the most famous chatbot today, being widely used as an interactive means of entertainment or for commercial processes.
Some still use it as a source of information, although it is not the best use, as mentioned. In fact, the company was recently sued for spreading a compromising lie. Find out more below.
OpenAI sued by radio host after ChatGPT lies
Recently, for the first time, the company was sued because of its AI, accused of defaming radio host Mark Walters, accusing him of embezzlement. The information came to light after journalist Fred Riehl asked for a summary of a lawsuit involving a pro-gun company in the United States, SAF (Second Amendment Foundation).
However, when doing so, the chat delivered a response full of untruths, saying that the radio host would have:
“[…] misappropriated funds for personal expenses without authorization or reimbursement, manipulated financial records and bank statements to conceal their activities.”
However, Mark has no ties to SAF. He hosts a show that shares the same ideals as the foundation, but his name is not on the actual lawsuit. That’s why Walters sued OpenAI for defamation.
Why and how did the robot invent this?
This type of behavior is called “hallucination” by AIs. It occurs because ChatGPT is unable to identify the difference between reality and fiction, and it also has to look for the best words to complete sentences logically. Therefore, you should not completely trust the information provided by the chat.
OpenAI founder Sam Altman himself has already admitted to this error in the software. In fact, it was not the first time it had happened. Politician Brian Hood was already accused by the AI of corruption in a bribery case involving the Australian central bank. However, Hood was responsible for reporting the crime, contrary to what was reported.
What can happen to the company?
Unfortunately, for those affected, the case will probably end there, as OpenAI would have to be held responsible for knowing that the information was false and acting in the same way.
However, according to experts, such as law professor Eugene Volokh, this is unlikely. Furthermore, those involved would have to prove that the accusations caused damage to their reputation.
Support our work ❤️
If you enjoyed this article, consider leaving a tip to help us keep publishing great content.
