I lead a book club once (Designing Data Intensive Applications)- read a chapter and meet every two weeks. Was a real flop. Attendance remained high, but only one other person actually finished the book.

What was a real slap in the face - maybe a year after that book had concluded, someone told me I should lead another one about this other topic. She had not finished the first book, and she wanted me to regurgitate another to the group?

Maybe there was value in the discussions that didn’t require the reading? But running seminars isn’t hobby material. If this was happening at my work, I may skim or not even read parts of the book, and still attend discussions.

There can be value, it's a networking opportunity and some companies could look favorably on it for promos.

I co-run a similar group at my company (added a comment above), what has worked for us and we realised this early on: not everyone will have the time to read and hence the sessions need to have a driver/lead. Assigning reading does not work unfortunately as many attend such groups voluntarily in addition to their primary work.

We as organisers (better to have at least 2) prepared before hand, either a presentation or a document, and then presented it in the group. While doing this, the group is free to discuss and interrupt at any time (we chose a slighly informal venue in the office).

Gradually, after about 10 sessions, we started seeing voluntary interest from the attendees to present a paper. And honestly this was an amazing feeling. So I'd suggest first finding a co-organiser who is interested in doing this and then pushing through the initial sessions by driving the topic yourself. That said, since you are preparing, you are free to choose the papers. We saw that if you choose papers related to a common larger theme that you are interested in, people would show up. Initially the attendance would be low, but with regular meetings, you'd start seeing regulars.