Is ChatGPT’s Hallucination Irreparable? Worries Raised by Researchers

Artificial intelligence (AI) chatbots like OpenAI’s ChatGPT, Microsoft’s Bing, and Google’s Bard have significant potential in helping people, but their universal technique, large language models, leads to generating an incorrect answer consistently. Hallucinations is the significant problem in AI chatbots because of their large language models that predict the most suitable answer based on internet data, but they lack the ability to determine whether the answer is factual. To address this issue, researchers from MIT proposed using multiple chatbots to create more than one answer to the same question and let them debate each other until one answer is determined the best. The researchers’ approach has improved chatbots becoming more factual while preventing them from making false claims. Despite AI’s exceptional capabilities, it must address the hallucination problem because it can further damage society, leading to online misinformation and deceit.