“AI safeguards” without instruments is just posture

France 24 reports that Donald Trump dismisses AI safety warnings as a “SICK conspiracy,” and then tells you that “tech leaders are calling for greater safeguards” as AI faces “increasing scrutiny” over cybersecurity, jobs, data centers, and national security.

That’s not policy coverage. It’s posture coverage wearing policy nouns.

A reader can argue with a policy. They can’t argue with “scrutiny.” “Safeguards” is not a safeguard until you name the instrument:

• Who are the “tech leaders” (names, organisations)?
• What safeguard are they actually advocating (licensing, evaluation mandates, liability rules, export controls, reporting requirements, something else)?
• Where is it meant to happen (Congress, an agency, a state, an international forum)?
• When does it take effect (a bill date, a regulatory timeline, an executive order, a court deadline)?

Without those handles, the piece trains you to experience AI governance as culture-war theater: one side sneers, the other side “calls for safeguards,” and the rest is a fog of stakes.

You can feel the importance. You just can’t point to what anyone is proposing to do.