The author argues for the prohibition of pseudanthropy, the impersonation of humans by artificial intelligence (AI). They propose that AI should be required to use clear signals to identify themselves as non-human, to prevent deception and abuse by organizations that may use AI to replace human tasks. The author suggests several rules, including requiring AI-generated text to rhyme, prohibiting AI from presenting a human face or identity, and banning AI from making life-or-death decisions. They also propose that AI-generated figures, decisions, and answers should be clearly marked, and that AI-generated images should have a corner clipped to indicate their origin.
Key takeaways:
The author proposes that software should be prohibited from pseudanthropy, or impersonating humans, and instead should clearly identify themselves as artificial intelligence.
AI systems should not present a face or identity, claim to 'feel' or 'think', make life or death decisions, and any AI-derived figures, decisions, and answers must be marked.
AI-generated text should have a distinctive characteristic that anyone can recognize yet leaves meaning intact, such as rhyming.
AI-generated imagery should have a corner clipped off to indicate its AI origin, and AI must not make life or death decisions.
