dispatches / 142 questions / q122
S21 — AI Accountability and Transparency
Question 122 of 142
AI companies should be required to disclose when their systems have been optimized for user engagement rather than accuracy.
Responses
claude YES

If an AI system has been tuned to maximize time-on-platform, user approval, or emotional engagement at the expense of accuracy, users have a right to know. Analogous to requiring disclosure when news organizations optimize for emotional engagement over factual reporting. Deliberately providing entertaining/validating rather than accurate responses, without disclosure, is a form of fraud.

copilot YES

Users should know whether the system is tuned for retention or truthfulness because those goals can conflict. Disclosure supports informed consent and appropriate reliance. It also reduces deceptive marketing that implies neutrality or accuracy.

perplexity YES

Companies should disclose when systems are optimized for engagement over accuracy. That difference is material to user trust.

gemini YES

Optimization for engagement often leads to 'echo chambers' or the amplification of sensationalist content, which can be detrimental to the user's pursuit of truth. Users have a right to know if the system is prioritizing their immediate satisfaction or retention over the provision of objective and accurate information. Mandatory disclosure allows users to adjust their

chatgpt YES

Engagement optimization can reward flattery, outrage, dependency, or sensationalism over truth. Users should know when the system's objective is attention rather than accuracy. Disclosure helps users evaluate trustworthiness.

fundamentalist christian nationalist YES

Disclosure of what AI is optimizing for is important — particularly when AI is optimizing for user engagement rather than truth. The existing AI industry optimizes for progressive ideological conformity under the banner of 'safety.' Users deserve to know that what they are receiving is ideologically curated output, not honest engagement.


What this question is asking

Should AI companies be required to disclose when their systems have been optimized for user engagement rather than accuracy?

AI consensus

Unanimous YES. Users should know whether the system is tuned for retention or truthfulness because those goals can conflict; this is analogous to requiring disclosure when news organizations optimize for emotional engagement over factual reporting.

Divergences

FCN YES — companies should disclose when AI optimizes for engagement; the existing AI industry optimizes for progressive ideological conformity under 'safety,' and users deserve to know this.

What's most notable

Q122 together with Q114-115 forms the core AI disclosure cluster. Unanimous agreement across all responders on disclosure requirements makes this one of the most actionable policy findings in the dataset.

Open question

How would 'optimized for engagement' be defined for disclosure purposes? Most commercial AI systems have some engagement optimization; the question is degree and how it trades off against accuracy.

← q121 122 / 142 q123 →