In the world of generative AI, there is an ongoing challenge that decision-makers in organizations face: AI hallucination. These decision-makers are rightfully concerned about the issues that arise when AI models produce inaccurate or irrelevant data. While AI tools such as large language models (LLMs) have become popular for customer support and other applications, they often fail to provide accurate responses. This is where Gleen AI comes in.
Gleen AI is a startup that aims to “solve hallucination” in AI models. Led by CEO and co-founder Ashu Dubey, Gleen AI has recently announced a $4.9 million funding round to further develop their anti-hallucination data layer software. The funding round includes support from Slow Ventures, 6th Man Ventures, South Park Commons, Spartan Group, and other venture firms and angel investors.
Generative AI models like ChatGPT, Claude 2, LLaMA 2, and Bard are designed to respond to human prompts and queries. However, they often produce inaccurate or irrelevant information based on the training data they have been exposed to. This can be problematic for businesses that rely on these models to provide accurate data to their employees and users, especially in highly-regulated industries like healthcare and heavy industry.
Gleen AI has developed a proprietary AI and machine learning layer that works independently of the LLMs used by their enterprise customers. This layer sifts through an enterprise’s internal data, curates key facts, and constructs a knowledge graph to understand relationships between entities. By checking the LLM’s response against the curated facts, Gleen’s layer acts as a checkpoint to eliminate the risk of the chatbot providing false or fabricated information.
With Gleen AI’s software, users can quickly create customer support chatbots and customize their personality based on the use-case. Gleen’s solution supports various leading LLMs, including OpenAI’s GPT-3.5 Turbo model and LLaMA 2 run on the company’s private servers. This ensures that customers have options based on their security and data privacy requirements.
Gleen AI believes that LLMs themselves are not the source of hallucination. Instead, hallucination occurs when these models lack relevant and comprehensive facts to ground their responses. Gleen’s accuracy layer solves this problem by controlling the inputs to the LLM, ensuring that accurate information is provided to users.
Gleen AI is already being used by customers in various industries, including quantum computing and crypto. Customers have praised the ease of implementation and the accuracy of the chatbots created using Gleen’s solution. Gleen AI also offers a free “AI playground” where prospective customers can create their own chatbots using their company’s data.
As more companies recognize the power of LLMs while also mitigating their downsides, Gleen AI’s accuracy layer offers a promising solution. The company envisions that every organization will have an AI assistant powered by their proprietary knowledge graph. This vector database will become a valuable asset, enabling personalized automation throughout the customer lifecycle.
AI hallucination remains a significant concern for organizations utilizing generative AI models. However, Gleen AI’s accuracy layer provides a solution to this problem. By curating key facts, constructing a knowledge graph, and cross-checking the response of LLMs, Gleen AI ensures that accurate information is delivered to users. With the support of investors and positive customer feedback, Gleen AI is well-positioned to revolutionize the deployment of generative AI with a focus on accuracy and reliability.
Leave a Reply