ChatGPT Nonetheless Not Assembly Data Accuracy Requirements, EU Data Safety Board Says

OpenAI’s efforts to supply much less factually false output from its ChatGPT chatbot usually are not sufficient to make sure full compliance with European Union knowledge guidelines, a job power on the EU’s privateness watchdog stated.

“Although the measures taken in order to comply with the transparency principle are beneficial to avoid misinterpretation of the output of ChatGPT, they are not sufficient to comply with the data accuracy principle,” the duty power stated in a report launched on its web site on Friday.

The physique that unites Europe’s nationwide privateness watchdogs arrange the duty power on ChatGPT final yr after nationwide regulators led by Italy’s authority raised considerations in regards to the broadly used synthetic intelligence service.

OpenAI didn’t instantly reply to a Reuters request for remark.

The numerous investigations launched by nationwide privateness watchdogs in some member states are nonetheless ongoing, the report stated, including it was subsequently not but doable to supply a full description of the outcomes. The findings had been to be understood as a ‘widespread denominator’ amongst nationwide authorities.

Data accuracy is likely one of the guiding rules of the EU’s set of knowledge safety guidelines.

“As a matter of fact, due to the probabilistic nature of the system, the current training approach leads to a model which may also produce biased or made up outputs”, the report stated.

“In addition, the outputs provided by ChatGPT are likely to be taken as factually accurate by end users, including information relating to individuals, regardless of their actual accuracy.”

© Thomson Reuters 2024
 

(This story has not been edited by NDTV employees and is auto-generated from a syndicated feed.)