Because half the world is below average.

S3 is just a general storage bucket, anything you could use a hard disk for, some app is using S3 for that instead. It's particularly common in serverless app backend architectures where no permanent storage exists at all, so volumes are not an option.

A good deal of hosting is done on ephemeral VM instances that get rebuilt, so you choose between volumes and S3, and S3 is a bit more flexible.

A fair chunk of modern corporate data ends up in OLAP systems that are now more often or not stored using s3 (or their MS/Google equivalents) in 'Data Lakes'. The concept of 'private data' whilst using cloud providers is an interesting one, but there has been enough work done by all parties involved to ensure that all but the most sensitive data is now created, stored and analysed using the systems provided by "Big Cloud".

Many applications use s3 as a data store for... Whatever. Uploaded or generated files, intermediate outputs, as a key-value store for large blobs.

You can put all sorts of things on s3. The PDF containing your bank statement, your medical test results, whatever - S3 is plenty secure enough for that with sane configuration, and even by default with sane account management.