Generative AI hallucinates up to 20% of the time, damaging trust with employees and customers. Four fixes: tie your AI to a real knowledge base, write specific prompts instead of yes/no questions, tell it to admit when it doesn't know, and be transparent about using AI.
February 6, 2024
Up to 20% of the time
Let's be honest. Is your AI keeping it real? Genai can boost your business, but it can also make mistakes. These are known as AI hallucinations. These can happen up to 20% of the time and cause trust issues between your employees and your customers. You definitely do not want that. So how do you prevent these hallucinations? Here's four tips. Number one Use a trusted LLM that's tied to your knowledge base for the most reliable results. Number two Craft specific AI prompts. Avoid yes or no questions. These limit the AI's ability to provide more detailed information. Remember to ask follow ups and provide tons of details. Number three Tell the LLM to literally be honest. Put in your prompt. If you don't know the answer, just say so. And finally, number four, be transparent about your AI use. Follow local laws and work with your legal advisors. And make sure you also have safeguards in place like toxicity and bias detection. After all, you want to protect your customers and your brand. Curious for more? Get a deeper dive in this blog post.
