Comment by AUF2026

Because that's the policy of ChatGPT. Disabling links helps protect user privacy, as sharing external links can lead to the potential sharing of personal data or exposure to harmful content. On the security side, preventing links helps mitigate the risk of directing users to phishing sites or malware-infected URLs. Preventing links also helps avoid external sources influencing responses, ensuring the information remains unbiased and solely reliant on the trained knowledge base.

All AI models are probabilistic models with weights, and if used improperly, they can cause more harm than good. I recommend using them with caution or, better yet, transitioning to DETERMINISM.