Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Medium bans AI-generated content from its paid Partner Program

bleepingcomputer.com

Post reply on HN
Print view View original

11–20 of 62 posts

Re: Medium bans AI-generated content from its paid Partner Program

#11

Post by lelandfe »

hk__2 wrote: post #7

> Any story using AI help must clearly mention this within the first two paragraphs. What’s AI here? Is the spelling correction tool in Google Docs "AI"? Are the suggestions from my iPhone "AI"? This is the kind of very vague sentence written by someone who doesn’t understand AI.

https://help.medium.com/hc/en-us/articles/22576852947223-Art...

> We define AI-generated writing as writing where the majority of the content has been created by an AI-writing program with little or no edits, improvements, fact-checking, or changes. This does not include AI writing tools such as AI outlining, or AI assisted fact, spelling, or grammar checkers

> Do I need to disclose the use of AI-assistance such as grammar or spell checkers, outline assistance, or fact verification?

> No, you do not need to disclose this use. You must disclose if you are using text or images generated by AI

Re: Medium bans AI-generated content from its paid Partner Program

#12

Post by EA-3167 »

felideon wrote: post #5

How is this policy enforceable? There aren’t any reliable ways to detect LLM-generated text, are there?

Read a TOS some time, it's enforceable because they can choose to do absolutely whatever the hell they want. If they're wrong, well... tough nuts.

Re: Medium bans AI-generated content from its paid Partner Program

#13

Post by Verdex »

rany_ wrote: post #6
felideon wrote: post #5

How is this policy enforceable? There aren’t any reliable ways to detect LLM-generated text, are there?

Using common sense, probably. You don't need an LLM detector to tell that something is AI generated.

It looks like you're getting downvoted, but my opinion is that your statement is correct.

Yes, technically, common sense doesn't work because some things written by AI won't be obviously written by AI and some things written by people can appear to be written by AI when it was not.

However, I suspect that medium doesn't actually care if something is written by AI or not. They want to stop the person writing hundreds of garbage articles a day. The person writing a nonstop deluge of poor quality articles sans AI is going to be erroneously banned. The person writing high quality articles 100% powered by AI won't be banned. AI will simply be the justification for culling what they do not want.

Re: Medium bans AI-generated content from its paid Partner Program

#15

Post by Verdex »

michaelbuckbee wrote: post #10

All publishers are grappling with this where they want generally higher quality content (defined as more engaging, intrinsic quality is something else) and less low effort spam or retreaded concepts. AI-generated content has really dropped the floor on how cheap it is to produce niche blogspam, and they're saying: "Hey if you're pushing out 1000s of articles a day on a topic and we look at it and it's insane gibberis…

I believe this to be the correct take away.

They don't actually want to ban AI content. They want to ban garbage content. If they claim that something is AI content, then it's not like there's a good defense a writer can offer. Short of meticulously recording themselves typing the article up on a typewriter in real life ahead of time. In real time so everyone can see them going through revisions.

This is AI and we're banning you, is simply a sufficiently socially acceptable reason to get rid of content that they otherwise don't want clogging up their site.

Re: Medium bans AI-generated content from its paid Partner Program

#16

Post by barfbagginus »

TuringNYC wrote: post #9
rany_ wrote: post #6

Earlier quoted context omitted.

Using common sense, probably. You don't need an LLM detector to tell that something is AI generated.

>> Using common sense, probably. You don't need an LLM detector to tell that something is AI generated. This could quickly turn into a witch hunt

Many people now believe things are AI generated the moment the writing says something they disagree with. That's fun!

Re: Medium bans AI-generated content from its paid Partner Program

#18

Post by overstay8930 »

rany_ wrote: post #6
felideon wrote: post #5

How is this policy enforceable? There aren’t any reliable ways to detect LLM-generated text, are there?

Using common sense, probably. You don't need an LLM detector to tell that something is AI generated.

this works until an entire generation of people are used to reading LLM generated text and they write like an AI

Re: Medium bans AI-generated content from its paid Partner Program

#19

Post by barfbagginus »

lelandfe wrote: post #11
hk__2 wrote: post #7

> Any story using AI help must clearly mention this within the first two paragraphs. What’s AI here? Is the spelling correction tool in Google Docs "AI"? Are the suggestions from my iPhone "AI"? This is the kind of very vague sentence written by someone who doesn’t understand AI.

https://help.medium.com/hc/en-us/articles/22576852947223-Art... > We define AI-generated writing as writing where the majority of the content has been created by an AI-writing program with little or no edits, improvements, fact-checking, or changes. This does not include AI writing tools such as AI outlining, or AI assisted fact, spelling, or grammar checkers > Do I need to disclose the use of AI-assistance such as g…

This is not very clear about the case where the text is AI generated and substantially edited, fact checked, and improved by a human co-author.

I think that kind of writing is perfectly valid. AI contexts can now include primary references, and with human critique, AI can fix a lot of its own style and factuality issues.

A good author writing this way will end up writing in their own voice, but spend less time drafting and revising, since they can delegate that to their AI assistant.

Re: Medium bans AI-generated content from its paid Partner Program

#20

Post by matwood »

Verdex wrote: post #13
rany_ wrote: post #6

Earlier quoted context omitted.

Using common sense, probably. You don't need an LLM detector to tell that something is AI generated.

It looks like you're getting downvoted, but my opinion is that your statement is correct. Yes, technically, common sense doesn't work because some things written by AI won't be obviously written by AI and some things written by people can appear to be written by AI when it was not. However, I suspect that medium doesn't actually care if something is written by AI or not. They want to stop the person writing hundreds…

Exactly. People are getting too caught up on AI or not. If an article is good, it's good, regardless of AI.
Post reply on HN