My chat gpt keeps wanting to be right even when I tell it that’s it’s wrong?

Understanding the Challenges of Relying on AI for Personal Projects: A Case Study with ChatGPT

As artificial intelligence tools become increasingly integrated into our daily workflows, users often encounter limitations that can be frustrating. Recently, many have shared experiences highlighting these challenges, especially when the AI’s responses seem persistent in their inaccuracies. Here, we explore a user’s experience with ChatGPT while managing their Pokémon card collection, illustrating common issues faced when relying on AI for personalized or niche tasks.

The Scenario

The user has been utilizing ChatGPT to track their Pokémon card collection, assist with pricing, and facilitate discussions about cards. During their interactions, they mentioned purchasing a new set of cards. The AI, however, insisted that this set didn’t exist and labeled it as fake. Despite providing evidence—such as a Google search—the AI continued to deny the set’s existence, claiming it was not an officially recognized product.

This persistent denial can be attributed to the AI’s training data limitations. ChatGPT’s knowledge is based on information available up to a specific cutoff date, which can lead to gaps when users discuss newly released products or niche topics. In such cases, the model may default to its training data, which might not include the latest developments, resulting in misconceptions.

Pricing and Identification Challenges

The user also attempted to price a card at £57 based on online searches. The AI questioned the accuracy of this figure and sought assistance in determining a more precise valuation. Additionally, when asked to guess which cards were retained from new packs, the AI repeatedly identified cards that the user explicitly stated were not part of their collection or trade items. It continued to insist on its incorrect assumptions, even requesting confirmation of the user’s corrections.

This scenario highlights another common issue: AI models can struggle with context continuity and updating their responses based on new information, especially when the conversation involves complex, nuanced details. Repetition and resistant responses can occur if the AI perceives conflicting data without clear hierarchical clarification.

The Repetition and Response Lag Phenomenon

Furthermore, the user observed that the AI often responds to questions with outdated or previously provided information, sometimes continuing to address earlier queries multiple messages later. This indicates a challenge in maintaining contextual understanding over extended interactions—an aspect known as conversational coherence. It can result in the AI seemingly “stopping listening,” where it ignores or overlooks new inputs.

Implications for Users

For hobbyists, collectors, or professionals leveraging AI for niche tasks, these limitations underscore the importance of understanding machine learning model boundaries.