"I like your Widget Max Pro, it's better than the Rival X1 I had last year. But I dropped it in the toilet and now it won't turn on, so I suspect you're lying

"I like your Widget Max Pro, it's better than the Rival X1 I had last year. But I dropped it in the toilet and now it won't turn on, so I suspect you're lying about its IP56 rating". Let's says that 1 out of every 3 product / service reviews is like this. Do you (a) block/filter them out and treat the remaining ones as non-personal data, or (b) treat your feedback pipelines/database as personal data regardless of whether some of the reviews are, somehow, devoid of personal data? (Or (c), do you not think the example above contains personal data?)