Study: AI Declined to Answer Prompts About Our Brand 31% of the Time (But Got Our Info Mostly Right When It Answered)

A recent study found that major AI platforms incorrectly answered questions about a specific company 31% of the time. While the AI often provided accurate information when it did respond, the refusal rate raises concerns about data reliability and access.
Key takeaways
- AI models failed to answer company-specific prompts frequently.
- Accuracy was good when AI provided an answer.
- Data access and completeness remain AI limitations.
- Verification is crucial for AI-generated information.
Why it matters
This research highlights potential blind spots in AI assistants' knowledge bases. Users should verify information, especially for niche or proprietary data, as AI may not always access or disclose relevant details, impacting decision-making accuracy.
Try this on SynaBot
Related AI assistants, prompts, and tools from the SynaBot catalog.
- Ordinary People PromptsOrdinary People Prompts provides meticulously crafted prompts to enhance AI interactions for individuals and teams seeking to generate, refine, and structure a wide range of content, from articles to emails.
- StudyableStudyable is an AI-powered education tool designed to tutor learners, generate practice materials, explain concepts, and streamline study workflows with features like AI summaries and custom flashcards.. Studyable revolutionizes learning with AI-driven tools for tutoring, practice generation, and progress tracking, off
- dotBRANDdotBRAND is an AI design tool that generates images, design assets, and brand visuals to streamline creative concept iteration for designers and marketing professionals.


