Auditing a 'Social Proof' Feature for Manipulation

Marketing wants to add a live-updating banner on the pricing page: "🔥 47 people are viewing this plan right now" and "Sarah from Chicago just upgraded to Pro 2 minutes ago." Engineering tells you privately that the "viewers" number is a randomized value between 20 and 60 that has no connection to actual traffic, and the "just upgraded" notifications are pulled from a static pool of 30 fictional names that cycle on a timer, not real purchase events.

1. Using the persuasive-vs-manipulative test from this subject, evaluate both banner elements. Would either pass if it were built on real data instead? Explain the difference.
2. Which named dark pattern category from Brignull's taxonomy does the fabricated version most resemble, and why?
3. What would you say in the meeting when marketing asks for this, and what would you propose instead?