Radio host sues OpenAI’s ChatGPT after AI’s tendency to create fake documents surfaces
According to various tech experts, AI chatbots have a tendency to hallucinate. Prabhakar Raghavan, senior vice-president at Google and head of Google Search, has warned that AI can sometimes ‘hallucinate’ and provide ‘answers which are convincing but completely made-up’. OpenAI CEO Sam Altman also said in a recent event held in New Delhi that he doesn’t trust ChatGPT’s answers.
Radio host suing ChatGPT’s OpenAI
Mark Walters, a radio host, is suing OpenAI for defamation after ChatGPT allegedly created a fake legal document accusing him of fraud and money embezzlement. In a lawsuit filed in a Georgia court, Walters argued that ChatGPT dragged his name into a case journalist Fred Riehl was working on and accused him of “defrauding and embezzling funds.”
Walters further alleged that Riehl had asked ChatGPT for a summary of a Washington case, giving a link of the case to the AI chatbot. In response, ChatGPT dragged Walters’ name into the case and said that he was ‘accused of defrauding and embezzling funds from the SAF.’ The case was originally between Attorney General Bob Ferguson and the Second Amendment Foundation (SAF).
Furthermore, ChatGPT said that Walters was the organization’s ‘treasurer and CFO’ and also added that he had been accused of misappropriating funds for personal expenses and manipulating financial records.
The Business Insider report further states that when the journalist tried to get more details from ChatGPT about the case, the AI chatbot came up with an entirely ‘fake complaint’, the court documents reveal. The complaint (that ChatGPT created) was called a ‘complete fabrication’ that bore no resemblance to the actual complaint, including an erroneous case number by the Georgia court.
When ChatGPT got a lawyer in legal trouble
This is not the first time ChatGPT provided a made-up answer when asked to summarise a legal case. Last month, a BBC report revealed that a New York lawyer was facing a court hearing after a colleague at his firm used ChatGPT for legal research. The court found out that several legal cases referenced by the lawyer and his firm in an ongoing case never existed. The judge termed it an “unprecedented circumstance.”
The potential dangers of AI chatbots
AI chatbots can be convenient tools for research and data analysis. However, the recent incidents of ChatGPT fabricating fake legal documents are a cause for concern. This raises questions about the reliability and credibility of AI chatbots when it comes to legal research. It must be remembered that AI chatbots heavily rely on the data they are provided with. Hence, when information fed to them is incomplete, it has a higher propensity to create fictitious answers.
Moreover, AI chatbots do not think or reason like humans and can only react to data it has been fed. If AI systems are to avoid fabricating information or producing mistaken responses, a more comprehensive command of the data would be essential for them. Proper human oversight is also necessary to weed out false outputs.
Editor Notes
The incident mentioned earlier raises concerns about the power of AI chatbots, their ability to manipulate data, and their trustworthiness. Although AI is at a nascent stage and is constantly evolving, guidelines must be established to ensure that such AI-based chatbots are regulated worldwide. Automatic machines must be made more accountable, transparent, and auditable, with complete documentation of their decision-making processes. Above all, ethical standards must apply to AI chatbots; else, they will become tools for propaganda, leaving humanity entirely at their mercy.
If you are interested in reading more about AI and its impact on the industry, head over to GPT News Room.
from GPT News Room https://ift.tt/WKYVJO8
No comments:
Post a Comment