Summary: AI21 Labs has released a new tool called the Contextual Answers API, which minimizes hallucinations and incorrect information generated by AI text generators. This tool restricts the text generator to answering queries based on the data provided by AI21’s client, ensuring more accurate responses.
Introduction: AI21 Labs has developed the Contextual Answers API to address the issue of hallucinated content and incorrect answers generated by AI text generators. These generators often provide imaginary answers when they cannot answer a question, leading to misinformation. The new tool improves the relevance and specificity of answers by relying on data uploaded by the organization using the software. It processes this data to provide natural language responses, without fabricating information.
Main Points: The Contextual Answers API distinguishes itself by solely using the provided data and not attempting to fill gaps with imaginary answers. This addresses the concerns of enterprises, as a recent KPMG survey revealed that 90% of organizations experimenting with generative AI are worried about hallucinated content. In addition to filtering hallucinations, Contextual Answers enhances generative AI chatbots by deducing misspelled or missing words based on context and improving grammar parsing.
Conclusion: The launch of Contextual Answers by AI21 Labs marks a significant step in combating generative AI hallucinations. With this tool, organizations using generative AI can ensure more accurate responses by restricting the text generator to the provided data. The company believes that as companies become smarter about implementing AI into their workflows, tools like Contextual Answers will play a key role in their AI strategies.