The chatbot repeated false claims 30% of the time and gave vague or not useful answers 53% of the time in response to news-related prompts, resulting in an 83% fail rate, according to a report ...
Some results have been hidden because they may be inaccessible to you