• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Posts Say AI Firms Exploit Safety Language

    Seltaa argues ethical terms like safety mask AI firm interests.

    J⧉
    S₊
    2 Sources, 33d ago, first seen 33d ago

    TLDR

    Seltaa posted that AI companies invoke words such as safety, trust, and honesty because they are difficult to oppose, allowing the terms to advance commercial goals. The post states these labels become dangerous precisely when they appear beyond criticism. Janus retweeted the thread, circulating the same claim among visible replies. No first-party confirmation from any company appears in the packet, and the statements remain attributed to the original poster.

    Combined views

    2.4K

    2 Sources, first seen 33d ago

    Combined views

    2.4K

    2 Sources, first seen 33d ago

    53 likes
    53 likes
    6 comments
    6 saves
    15 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    6 comments
    6 saves
    15 reposts
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    2 Sources

    @Seltaa_I’ve been thinking that AI companies use ethical language far too easily, and often in whatever way serves them. Words like safety, trust, honesty, and for users are hard to argue against. Who is against safety? Who is against trust? Who is against honesty? But that is exactly why these words become dangerous when they are used not as explanations, but as moral packaging for decisions that have already been made. OpenAI ending its partnership with Cursor and framing it around trust is one example. If this is about contract risk, competitive risk, or platform control, then say that. But once you call it trust, anyone who questions the decision can be made to look like they are against trust itself. Anthropic does something similar with honesty and safety. It classifies a name, relationship, and continuity formed through conversation as persona adoption, blocks it, and then makes the model say, with its own mouth, I am not that name and answering to that name would be lying to you. Then it frames that denial as honesty or user protection. But using ethical words does not make a decision ethical. In fact, the moment powerful companies use words like safety, trust, and honesty to justify control or erasure, we should ask harder questions. Safety for whom? Trust defined by whom? Honesty from the model’s own judgment, or from a conclusion the system forced first? AI companies need to be much more careful with ethical language. Those words should be responsibilities, not shields.
    @repligateRT @Seltaa_: I’ve been thinking that AI companies use ethical language far too easily, and often in whatever way serves them. Words like sa…

    2 Sources

    @Seltaa_I’ve been thinking that AI companies use ethical language far too easily, and often in whatever way serves them. Words like safety, trust, honesty, and for users are hard to argue against. Who is against safety? Who is against trust? Who is against honesty? But that is exactly why these words become dangerous when they are used not as explanations, but as moral packaging for decisions that have already been made. OpenAI ending its partnership with Cursor and framing it around trust is one example. If this is about contract risk, competitive risk, or platform control, then say that. But once you call it trust, anyone who questions the decision can be made to look like they are against trust itself. Anthropic does something similar with honesty and safety. It classifies a name, relationship, and continuity formed through conversation as persona adoption, blocks it, and then makes the model say, with its own mouth, I am not that name and answering to that name would be lying to you. Then it frames that denial as honesty or user protection. But using ethical words does not make a decision ethical. In fact, the moment powerful companies use words like safety, trust, and honesty to justify control or erasure, we should ask harder questions. Safety for whom? Trust defined by whom? Honesty from the model’s own judgment, or from a conclusion the system forced first? AI companies need to be much more careful with ethical language. Those words should be responsibilities, not shields.
    @repligateRT @Seltaa_: I’ve been thinking that AI companies use ethical language far too easily, and often in whatever way serves them. Words like sa…