Apropos automated approval of "tone". The company I work for now has a policy that documents must not use any "potentially biased" terms (e.g. "whitelist" or "master"). We now have automated agents that do simple search-and-replace (e.g. s/whitelist/allowlist/) over our internal Wiki pages. With no intention of debating the value or politics of such a policy, I wonder whether anyone is contemplating employing ChatGPT for more pervasive automated tone-control.