Mark Penn: Five fixes to save a billion and a half hours

Penn's argument is that the digital friction we endure — consent banners, multi-step logins, CAPTCHAs, mandatory registrations and sprawling terms — was not designed for users. It accumulated as companies responded defensively to lawsuits, fines and a growing pile of privacy and security regulations. Because each company solved the problem on its own, the result is a patchwork of buttons and banners replicated across every site and app.

He points to Apple's passkeys as an example of a better path, and imagines an AI plugin that fills in forms automatically. He also proposes an independent rating system for terms and conditions, reasoning that average people cannot give meaningful consent to documents they cannot read. Underlying these ideas are real concerns about tracking, security and legal liability, which he acknowledges.

The stakes hinge on whether lawyers and regulators accept simpler alternatives, or whether the same defensive habits follow AI assistants into everyday use. Penn's warning is that without intervention, the clutter could spread to ChatGPT and Claude, turning simple questions into sign-in screens and liability waivers.