Talk:Louis Rossmann - Video Directory

Hi there: I have never used wikipedia as an editor before and was wondering which is the best way to contribute to this project while i learn the web basics. Will translations into spanish be needed at some point?

Good morning, there are numerous guides on the mediawiki and Wikipedia websites that provide an introduction to how to use wiki. My recommendation would be to learn some of the basics from those systems. If you’ve done something wrong with an article, staff will come through and edit it for you or provide additional context. We’re not going to remove content unless it’s blatant vandalism or a violation of the rules. Feel safe in knowing that any contribution you do make will be used. Thanks for reaching out and for asking for help. If there’s any specific questions you have I’d be happy to help and I know any other member of the team would as well. - Atsumari (talk) 13:22, 16 January 2025 (UTC)

I think that this video directory could benefit from some AI sentiment tools to make it easier for us to filter through results. I understand the general hesitance for using tools like this but I'd like to lay out my plan to get approval from the community before attempting some large changes.

1. Using the undocumented youtube transcript api, format the entire script of each video into context and ask the AI to summarize them with emphasis on Consumer Protection, likely using the same language and emphasis as our welcome guide.
2. Take this summary and place it into the context, this time asking for what kinds of consumer protection applies, provide specific examples so it is more likely to tokenize tags
3. Take this summary and place it into the context, this time asking whether it thinks this is a consumer protection video
4. Take this summary and ask it to identify key identities, such as corporations, people, and so on
5. Do some data normalization on extracted tags to make sure they are normalized
6. Reformat current tables to include this table to allow people to sort by Consumer Protection Topic, Identity, or Likely Consumer Protection Video. Additionally, provide summary to help people decide for themselves if this is content they want to write a full article about.

Please let me know what the community thinks of this, I think this would be useful to the expansion of these Videos, without creating AI crud. I will work on this script regardless and will make source available. FierclyIndependant (talk) 01:10, 17 January 2025 (UTC)

Hey All,

I wanted to get a basic prototype working. It looks like I may have one more goal completed here: This also collects the transcript as part of the process. Currently, no punctuation is added but there is the code to add it using another model. Please see this GitHub for example code, I will finish up the proof of concept tomorrow after work for the community to review. [1] FierclyIndependant (talk) 03:59, 17 January 2025 (UTC)