A working group under the EU's privacy watchdog has said that OpenAI's efforts to reduce the number of false answers from their ChatGPT chatbot are insufficient to achieve full compliance with EU data processing rules, UNN reports citing Reuters.
While the measures taken to comply with the principle of transparency are useful for avoiding misinterpretation of ChatGPT results, they are not sufficient to comply with the principle of data accuracy
Last year, following concerns from national regulators, led by the Italian authorities, about the widely used artificial intelligence service, the European body that coordinates national privacy authorities set up a working group on ChatGPT.
"In fact, due to the probabilistic nature of the system, the current training approach results in a model that may also produce biased or fictitious results. In addition, the results provided by ChatGPT are likely to be perceived by end users as factually accurate, including information relating to individuals, regardless of their actual accuracy," the report says.
Recall
OpenAI has signed an agreementwith News Corp (an American multinational media holding company) with a probable value of more than $250 million. The deal will allow ChatGPT to learn how to work with the archives of a number of well-known media outlets. AI will have access to copyrighted content.