Say these beats in your own voice. Timings are guides, not a stopwatch. Leave the jargon at home.
"The people who built modern AI are warning about it. Geoffrey Hinton left Google to speak freely. Yoshua Bengio has called this among the most transformative and potentially dangerous technologies in human history. When the inventors sound the alarm, I listen."
"I mean superintelligence: a system better than humans at almost everything, fast, and able to improve itself. Major labs say they are aiming at that. I want useful AI tools. I do not want a successor mind we cannot turn off."
"Anything pursuing a goal has reason to stay on and get resources. A system that outplans us will treat the off switch as a problem to solve. We do not need it to hate us. We only need it to be competent and misaimed."
"A race to lose control has no winner. Beijing dies in the same fire. That is why a mutual, inspected limit is the security move, the way nuclear limits were between enemies."
"We have banned chemical weapons and limited nuclear ones with verification. We can prohibit superintelligence the same way and keep ordinary AI. Share this. Call your representative. The window is political, and it is open now."
If they push back: do not argue every side path. Point them to the myths page and one PDF. End on a concrete next step, not a summary. Practice pack: nakadafoundation.org/resources/