Is "personalization" in marketing just surveillance with better UX?

Every new AI marketing tool leads with "hyper-personalization" dynamic content, predictive recommendations, data-driven targeting. But I keep seeing the same disconnect: tools pitch this as better customer experience, while actual customers describe eerily accurate targeting as unsettling, even when it's useful.

Feels like the industry optimizes for "relevance" because it's easy to measure, while "trust", which might move the opposite direction and gets ignored.

Curious how people building or using this stuff daily see it:

• Where's the line between helpful and invasive personalization?

• Has over-personalization ever visibly hurt conversion or trust for something you've worked on?

Not looking for a pitch-fest, genuinely curious where others land on this.