AI is not alive. It is not plotting against us. It does not decide who to silence or who to promote. It simply reflects the people who build it.
That point became clear to me again when I tried to make a set of graphics about open discourse. I chose a line I love: “We need more conversations that make people uncomfortable. That is how you grow.” My goal was simple, put those words next to the speaker in a clean, respectful design. The AI tool blocked me every time.
That moment reminded me that AI is not the one making these choices. People are.
Find out what's happening in Long Islandfor free with the latest updates from Patch.
AI does not wake up and decide what is acceptable. It runs the rules written by developers, shaped by policy teams, and signed off by lawyers. These guardrails are meant to protect, but they also stop lawful and constructive conversations. Sometimes they do both at once.
This is why I see AI as a mirror of human failure. It reflects the values we encode into it, the blind spots we ignore, and the fears we build into our systems. The danger is not that AI becomes sentient but that it scales the choices we made without question.
Find out what's happening in Long Islandfor free with the latest updates from Patch.
Edelman’s 2024 Trust Barometer shows sixty one percent of people only trust AI if they know how it works. That is the path forward. Trust is built with transparency. When a tool flags something I write, it should explain why and show me how to fix it. That turns a block into a learning moment.
Anthropic has taken a step in this direction with its constitutional AI approach, training models on a written set of principles and publishing them for public review. It feels like pulling back the curtain, seeing the logic behind decisions and knowing where to push for better ones.
AI is not the censor. It is the code we wrote and the choices we made. If we want AI to allow hard conversations, we have to demand that from the people building it. When a tool says no, that moment should begin the dialogue, not end it. Friction is where democracy begins.
Where have you seen AI say no in a way that felt more like a human decision than a machine rule? Share it. This is the conversation we need if we want better tools, better rules, and better outcomes.
