3 ms·

This is a good write up on the perils of messing with a fine-tuned mechanism that is Section 230 and potential devastating effects it could have on the internet

by lovecg 5y ago

This is a good write up on the perils of messing with a fine-tuned mechanism that is Section 230 and potential devastating effects it could have on the internet as we know it. If not carefully drafted, any such bill will open up internet companies to many many lawsuits that can no longer be easily dismissed. The article goes into some examples where not getting the language quite right can blow the whole thing up. I also found the following article a good starting point for learning about what Section 230 is and is not: https://www.techdirt.com/articles/20200531/23325444617/hello-youve-been-referred-here-because-youre-wrong-about-section-230-communications-decency-act.shtml https://www.techdirt.com/articles/20200531/23325444617/hello...

donmcronald 5y ago
> The bill eliminates Section 230 for making a personalized recommendation of information that materially contributes to a physical or severe emotional injury. IMHO that’s purposely worded to limit the bypass to people, not companies. That’s why the SV rep supports it. It’ll allow people with no money or chance of winning to sue tech companies and that’s it. Paraphrasing someone that commented on HN once and I totally agree, services providing reverse chronological feeds should be considered platforms while services providing algorithmic feeds should be considered publishers. There should be no qualifier. Anything that’s a personalized feed should be considered publishing and companies should be liable for promoting that content.
- krapp 5y ago
  >services providing reverse chronological feeds should be considered platforms while services providing algorithmic feeds should be considered publishers A reverse chronological feed is no more or less algorithmic than any other - any sort requires an algorithm. You're just criminalizing algorithmic complexity, which is insane. There's no reason the same n items in a list should expose a site owner to different degrees of legal liability because they're sorted one way versus another. What about cases when the user can choose a different sort? Is a company liable based on whether or not a user decided they wanted to see content in anything other than strict chronological order? Is Twitter publishing or promoting content as soon as users tag it? Does sorting by karma make Hacker News a "publisher" of content? Do downvotes and flags count as "promotion" of the downvoted content, or all non-downvoted content?
  - donmcronald 5y ago
    > What about cases when the user can choose a different sort? Is a company liable based on whether or not a user decided they wanted to see content in anything other than strict chronological order? No. If I’m asking for it and they give me what I asked for, that’s not publishing. If they pick some content that would normally be buried and jam it in my face because some ML algorithm thinks it’ll cause engagement that’s publishing. > You're just criminalizing algorithmic complexity, which is insane. I don’t think it’s insane. I think letting black box ML algorithms moderate public discourse is insane. I have no problem with that being turned into a liability for big tech.
    - adgjlsfhk1 5y ago
      what about spam filters?
- lovecg 5y ago
  The proposed bill would not make this distinction though. Both of these would fall under the “personalized recommendation” umbrella, and so would Gmail and a ton of other services.

by lovecg 5y ago

donmcronald 5y ago

> The bill eliminates Section 230 for making a personalized recommendation of information that materially contributes to a physical or severe emotional injury. IMHO that’s purposely worded to limit the bypass to people, not companies. That’s why the SV rep supports it. It’ll allow people with no money or chance of winning to sue tech companies and that’s it. Paraphrasing someone that commented on HN once and I totally agree, services providing reverse chronological feeds should be considered platforms while services providing algorithmic feeds should be considered publishers. There should be no qualifier. Anything that’s a personalized feed should be considered publishing and companies should be liable for promoting that content.

krapp 5y ago
>services providing reverse chronological feeds should be considered platforms while services providing algorithmic feeds should be considered publishers A reverse chronological feed is no more or less algorithmic than any other - any sort requires an algorithm. You're just criminalizing algorithmic complexity, which is insane. There's no reason the same n items in a list should expose a site owner to different degrees of legal liability because they're sorted one way versus another. What about cases when the user can choose a different sort? Is a company liable based on whether or not a user decided they wanted to see content in anything other than strict chronological order? Is Twitter publishing or promoting content as soon as users tag it? Does sorting by karma make Hacker News a "publisher" of content? Do downvotes and flags count as "promotion" of the downvoted content, or all non-downvoted content?
- donmcronald 5y ago
  > What about cases when the user can choose a different sort? Is a company liable based on whether or not a user decided they wanted to see content in anything other than strict chronological order? No. If I’m asking for it and they give me what I asked for, that’s not publishing. If they pick some content that would normally be buried and jam it in my face because some ML algorithm thinks it’ll cause engagement that’s publishing. > You're just criminalizing algorithmic complexity, which is insane. I don’t think it’s insane. I think letting black box ML algorithms moderate public discourse is insane. I have no problem with that being turned into a liability for big tech.
  - adgjlsfhk1 5y ago
    what about spam filters?
lovecg 5y ago
The proposed bill would not make this distinction though. Both of these would fall under the “personalized recommendation” umbrella, and so would Gmail and a ton of other services.