05/06/2026
Recent studies show that AI tools can ‘hallucinate’ as often as 10% of the time!
And it will be confident in its misinformation. AI tools can be incredibly useful, but they are not automatically accurate.
One of the biggest risks is that AI can generate misinformation and present it as fact. That’s especially risky when you’re working on grant applications, evaluation reports, policy documents, or public-facing communications.
Use this prompt, or AI tool settings:
“Use evidence-based information only. Do not fill in gaps. Include placeholders for any missing information.”
This reduces the risk of AI inventing details, overstating claims, or adding unsupported information.
At NKC, we can help your organization develop AI use protocols and guardrails so your team can use AI more confidently, effectively, and safely.