AI Will Never Be Ethical or Safe

This article argues that AI will never be entirely ethical or safe because ethical and safe conduct depend on context and intent, which cannot be reliably known or verified. It uses examples like prompting dangerous information or using a firearm to illustrate how context and intent change meaning. The author contends that AI inherits the same social contract as human interaction, where trust is assumed but fragile. Even AI companies like Anthropic acknowledge this challenge but cannot solve it, as users often omit or lie about their context and intent. The piece concludes that AI is a tool that can be used ethically or unethically, and attempts to filter based on intent are naive since most people do not announce their true intentions.