Skip Navigation
Community Promo @lemmy.ca
πŸ‡¦πŸ‡Ίπ•„π•¦π•Ÿπ•₯π•–π••π•”π•£π• π•”π••π•šπ•π•– @lemm.ee

News summary for AI summarised news articles

Community focused on world news with ai summaries to allowing you to stay up to date without as heigh of a time investment.

Check out !news_summary@hilariouschaos.com

I'm welcoming to all political opinions and viewpoints (just engage in good faith and don't call for actionable violence) i try to avoid taking mod action against users unless absolutely nessasary.

I'm seeing a lot of new users coming from reddit. I would if you are new and want a place to keep up to date and express ur views in a place that won't censor you please join.

4 comments

  • How do you deal with AI summaries distorting news sources by hallucinating quotes or having factual errors about articles?

    Like Apple AI headlines: https://www.bbc.com/news/articles/cq5ggew08eyo

    • I use a finetuned T5 summarisation model that is relatively accurate. It has some minor issues with occasional miss assigning quotes but it doesn't hallucinate like a traditional GPT style model does. It is 60% identical to that of a human summary and >95% accurate in terms of meaning. It is more accurate than traditional nonai based summarisstion tools (I'm not sure how it compares to a human) but I belive it is as accurate and nonbias as possible.

      Its biggest flaw is actually the traditional nonai web scraper which sometimes pulls the wrong content. Its all foss so if u wanna go make a pull to improve it that would be greatly appreciated.

      EDIT: I've been experimenting with having a tradition GPT LLM look over the summary and original to catch these errors but have had little to no success without using large models which I cannot run on my local hardware (I unfortunately can't afford to pay for inference at the scale my bot runs).

      • Thanks for the explanation. I think if you combined that with a method to retract or edit summaries based on human reports, you can probably fill in the remaining 5%. I am unsure how feasible that would be though. Good luck with the community!