OpenAI states that allowing GPTBot access to websites can enhance AI models’ accuracy, capabilities, and safety. However, while this benefits OpenAI, website owners that have had their data scraped by Chatgpt receive no rewards.

AI Chatbots like ChatGPT acquire knowledge by gathering information from websites, which they later reproduce as answers without indicating the sources. Google’s Bard AI chatbot began including citations in May 2023, following Microsoft’s Bing Chat. In contrast, ChatGPT does not provide citations.

For instance, when questioned about recent innovations in solid-state batteries for electric cars, ChatGPT provides various examples without specifying any sources, including academic journals or publications reporting on the research. An experiment conducted at Indiana University revealed that ChatGPT may overlook or misquote citations even when provided with sufficient information in the prompt.

The fundamental issue here is that users who rely on ChatGPT for answers may no longer feel the need to conduct a traditional Google search, visit different websites, or find answers independently. As a result, publishers experience a decline in essential web traffic that sustains their livelihoods. Moreover, there is a risk of incorrect information being disseminated if ChatGPT provides inaccurate responses.

The post AI Chatbots and the Issue of Source Attribution appeared first on TS2 SPACE.