The Autonomous Conversational AI startup Got It AI recently unveiled a ground-breaking new “Truth Checker” AI that can spot when ChatGPT is fabricating answers to user requests while responding to them throughout a wide collection of articles or knowledge bases. With the help of this invention, ChatGPT-like experiences can be implemented without running the danger of giving customers or staff members the wrong information. Now that generative conversational AIs have access to massive knowledge bases, such those used for internal user support questions or for external customer assistance, businesses may confidently use them.
The Truth Checker AI trains itself autonomously for one task—truth checking—using a separate, cutting-edge Large Language Model (LLM) based AI system and a target domain of information (such as a sizable knowledge base or a collection of articles). With the same content, ChatGPT or an underlying LLM like a GPT-3.5 model can be used to respond to queries in a contextual, multi-turn chat dialogue. Before being displayed to the user, each response is examined for veracity. When an erroneous response is found, the user is not shown the response. Instead, a link to pertinent publications that include the solution is offered.
“We tested our technology with a dataset of 1000+ articles across multiple knowledge bases using multi-turn conversations with complex linguistic structures such as co-reference, context, and topic switches”, “ChatGPT produced incorrect responses for about 20% of the queries when given all the relevant content for the query in its prompt space. Our Truth Checker AI was able to detect 90% of the incorrect responses, without human help. We will also provide the customer with a simple user interface to the Truth Checking AI, to further optimize it, identify the remaining inaccuracies and eliminate virtually all inaccurate responses.”
Chandra Khatri, former Alexa Prize team leader and co-founder of Got It AI
“Our technology is a major breakthrough in autonomous conversational AI for ‘known’ domains of content, such as enterprise knowledge bases, versus ‘open domain’ such as the entire world wide web”, “It goes beyond prompt engineering, fine tuning, or just a UI layer. It is a proprietary model that enables us to deliver scalable, accurate and fluid conversational AI for customers planning to leverage generative LLMs. Truth checking the generated responses cost-effectively, is the key capability that closes the gap between an R&D system and an enterprise ready system.”
Amol Kelkar, formerly an architect for Microsoft Office Online and co-founder of Got It AI