According to TechCrunch AI, Microsoft’s terms of service for Copilot state that the AI assistant is “for entertainment purposes only,” a disclaimer that aligns with warnings from AI skeptics about the reliability of AI model outputs.
The report notes that AI companies themselves are cautioning users against placing blind trust in their models’ responses through their terms of service agreements. This represents a significant acknowledgment from Microsoft about the limitations of its AI technology, even as the company actively promotes Copilot’s capabilities to consumers and enterprise customers.
The disclosure highlights a growing pattern among AI providers of including liability-limiting language in their service agreements, effectively shifting responsibility for AI-generated content accuracy onto users. This approach underscores the ongoing challenges in AI development around ensuring consistent reliability and factual accuracy in large language model outputs.