MR Forums used in Google dataset for training AI’s

[retta283]

It is not surprising to me.

When Google is the one doing it, there is likely little that can be done to object, but it does not sit well with me. Of course what we say here is totally on public record, and sites like the Web Archive may be picking it up, but to nab it up and use it to build a commercial product on other people's work is unfortunate. Nothing new in the Internet age, this is a fairly shameless example and perhaps it is making the job of creating these AI easier.

We are all data sets now. It is an interesting thought that we are being used to create a new form of intelligence, I didn't suspect the internet would be so intertwined in the creation of these technologies 25 years ago. I never thought I'd even live to see what we are beginning to see.

 

[PhoenixDown]

The EU tends to be more privacy focused than the US. I wonder what their regulatory approach to all of this will be.

Ideally, in a situation like Google's, the spiders that crawl the site for searches are different from the spiders used to train their AI. If that's the case, then it should be possible to block one vs the other. Or if its the same spider, then it should be possible to use a tag to deny it for AI.

The copyright aspect is a larger issue and as old as the internet. I am for fair purpose (like Internet Archive). I understand that by posting on a forum, my content may be used for marketing and advertising. However can those right be arbitrarily assigned 2 - 3 times over to these AI'? Not sure.

 

[laptech]

The EU tends to be more privacy focused than the US. I wonder what their regulatory approach to all of this will be.

Ideally, in a situation like Google's, the spiders that crawl the site for searches are different from the spiders used to train their AI. If that's the case, then it should be possible to block one vs the other. Or if its the same spider, then it should be possible to use a tag to deny it for AI.

The copyright aspect is a larger issue and as old as the internet. I am for fair purpose (like Internet Archive). I understand that by posting on a forum, my content may be used for marketing and advertising. However can those right be arbitrarily assigned 2 - 3 times over to these AI'? Not sure.

The posts in here are our works, the copyright of those posts belong to us, the writer BUT we abdicate our rights of copyright to MR by way of the sign up rules. When we make an account we agree to hand over rights of our works to them. MR in it's terms and conditions will have rules on what 3rd parties can do with the posts in here. Question for us members now is did Google approach MR for it's permission to use the site for it's training of the AI and did MR give them that permission.

 

[KaliYoni]

how is this any different then what google does/track and monetize now?

To me, it's like the difference between a photocopier and a typewriter.

 

[HDFan]

Because MR is heavily biased towards everything Apple does this mean every time someone asks the AI an Apple related question it is always going to respond with a favorable response towards Apple?

But any bias would likely be balanced by a PC site in the dataset.

the spiders that crawl the site for searches are different from the spiders used to train their AI.

Seems like a good suggestion. But there is no one in control who could make it happen.

 

[maflynn]

Ideally, in a situation like Google's, the spiders that crawl the site for searches are different from the spiders used to train their AI. If that's the case, then it should be possible to block one vs the other. Or if its the same spider, then it should be possible to use a tag to deny it for AI.

But its in Google's best interest not to make them any different then indexing ones. They want to make sure you don't block them.

 

[laptech]

I see there has been no response from MR over this which implies to me they gave Google permission to have our data. I wonder how much MR was paid to hand over our data.

 

[maflynn]

how much MR was paid to hand over our data.

Probably the same amount Google pays people when they use gmail or their other services - nothing.

Maybe a minor point but technically its now MacRumor's data because you waived all rights to your content when signing up.

You warrant and acknowledge:

• that you give MacRumors permission to store your posts in a database, display them, and reproduce them on a website and by other media.
• that you will not post any messages that are obscene, vulgar, hateful, abusive, threatening, sexually explicit, or that violate any law.
• that you will observe all Forum Rules, including the rules that prohibit posts with personal insults, advertising, or the promotion of a product, service, organization, or website, except where explicitly permitted.
• that you understand that your posts and the user name you select will be visible to the Internet and to search engines, that you are responsible for your own privacy in deciding what to post, and that your posts will remain visible even if you leave MacRumors or your account is closed for any reason.
• MacRumors may re-purpose and re-post your comments and related content and images elsewhere on its forums, MacRumors web sites, and on social media accounts.
• that your membership may be temporarily suspended or permanently banned for violation of the Forum Rules.
• that you will be permanently banned if you re-register while your membership is suspended.

The owners of MacRumors reserve the right to:

• remove, edit, move, or close any discussion or message for any reason.
• close memberships that are detrimental to the forum community.
• re-purpose and re-post your comments and related content and images elsewhere on its forums, MacRumors web sites, and on social media accounts
• make changes to the Forum Rules without notice.

 

[Scepticalscribe]

Probably the same amount Google pays people when they use gmail or their other services - nothing.

Maybe a minor point but technically its now MacRumor's data because you waived all rights to your content when signing up.

Thanks for posting this, a timely reminder, indeed, to what we signed up to when we joined the forum.

I had forgotten much of this.

I am aware that you are no longer a member of the staff, but I wonder whether anyone (perhaps on the current staff) can shed any light on the meaning of the verb "re-purpose" when used in the conext cited here?

 

[maflynn]

can shed any light on the meaning of the verb "re-purpose" when used in the conext cited here?

Even when I was a mod, this sort of thing hadn't come up, so in my ignorance, I would imagine the repurpose is that they'll use the posts to create stories. For instance there's been news stories about an issue where they point to a thread or specific posts.

Ultimately, it may be just standard language to give the site owner the freedom, but Arn really hasn't exercised that freedom.

 

[Scepticalscribe]

Even when I was a mod, this sort of thing hadn't come up, so in my ignorance, I would imagine the repurpose is that they'll use the posts to create stories. For instance there's been news stories about an issue where they point to a thread or specific posts.

Ultimately, it may be just standard language to give the site owner the freedom, but Arn really hasn't exercised that freedom.

Thank you for your reply. Much appreciated.

I'll readily admit that I had never closely examined this material.

Initially, all that concerned me were the "don'ts", - for, this is the first online forum that I had ever joined, and this whole world was entirely new to me - the stuff not to do, or say, or write, to stay within forum rules and regulations.

Candidly, it never occurred to me to closely read, or mull over, or give thought to, the rest of this material, or that it would have any real impact on how one might choose to post, or exist, in the online space, or on an online platform; I had assumed that it was simply akin to those contracts with exclusion clauses and paid it little further heed.

However, if anyone else would care to clarify the meaning of the verb "re-purpose" in the context of the material that @maflynn has cited, I would be both interested to read it, and grateful for some clarification.

 

[laptech]

Google has changed it's privacy policy to say that it is going to collect information from 'publicly available resources' to train it's AI bot. MR is publicly available, you do not need a login to view some sections of the site. So this mean then what ever members post can be harvested by Google's AI bot. Where does MR stand on this? Do members of this site have a say as to what Google is doing? Or is MR just going to allow Google to harvest what ever they want from MR?

Source:

Now anything you post publicly online could be used to train Google's AI

Post at your own risk.

metro.co.uk

 

[I7guy]

Once it’s on the internet it’s forever.

 

[MacMan2013]

Google can harvest this: someday someone will start a class action lawsuit against you for IP infringement.

 

[maflynn]

Or is MR just going to allow Google

What do you suggest in preventing google from harvesting the information?

 

[adrianlondon]

What do you suggest in preventing google from harvesting the information?

I suggest every forum member just continues to write rubbish so that the data has no value.

 

[maflynn]

I suggest every forum member just continues to write rubbish so that the data has no value.

How will that be any different then what is typically posted

 

[adrianlondon]

How will that be any different then what is typically posted

Exactly. Hence "continues". And that includes however many posts I've made!

 

[adrianlondon]

Google can harvest this: someday someone will start a class action lawsuit against you for IP infringement.

Authors file a lawsuit against OpenAI for unlawfully ‘ingesting’ their books

Mona Awad and Paul Tremblay allege that their books, which are copyrighted, were ‘used to train’ ChatGPT because the chatbot generated ‘very accurate summaries’ of the works

www.theguardian.com

It'll be interesting to see how this goes, and whether it sets a precedent.

 

[MacMan2013]

Authors file a lawsuit against OpenAI for unlawfully ‘ingesting’ their books

Mona Awad and Paul Tremblay allege that their books, which are copyrighted, were ‘used to train’ ChatGPT because the chatbot generated ‘very accurate summaries’ of the works

www.theguardian.com

It'll be interesting to see how this goes, and whether it sets a precedent.

Yes that article caught my attention too. Appears to be in the Federal District for Central California, but would undoubtedly be appealed so could set a precedent across the US. Interesting comment that England has no fair use doctrine so the chances of an AI developer losing to a content creator would probably be even stronger.

 

[KaliYoni]

Do members of this site have a say as to what Google is doing? Or is MR just going to allow Google to harvest what ever they want from MR?

My guess–I don't have any insider knowledge of MR or Google–is that MR cannot block GOOG crawls of MR because it would affect MR's use of affiliate links, ad networks, and SEO tactics. I wouldn't be surprised if GOOG compels a high level of access for any website that relies on Google for revenues, backend services, or page views.

Having said that, if a webmaster or individual user is sufficiently motivated, it is possible to reduce GOOG surveillance somewhat. For example:

(remove your Wi-FI network from scanning by Google Maps cars and other Google operations)

Control access point inclusion in Google's Location services - Google Maps Help

To improve Location services and estimate the location of a device, Google uses publicly broadcast Wi-Fi information from wireless access points and GPS, cell tower, and sensor data. This data is limi

support.google.com

(suppress crawl information from appearing in Google Search results)

Block Search Indexing with noindex | Google Search Central | Documentation | Google for Developers

A noindex tag can block Google from indexing a page so that it won't appear in Search results. Learn how to implement noindex tags with this guide.

developers.google.com

But I can't find anything discussing if similar options are provided by GOOG's AI teams. GOOG makes these options difficult to find (or even know about) by not publicizing them and by burying them deep in its website. Worse, the Google AI homepage and site is all PR twaddle.

 

[maflynn]

Once it’s on the internet it’s forever.

Its ironic - at least to me, that some people incur all sorts of self righteous anger over this but happily post on social platforms. If you interact with social media, be it a fan site like MR, or facebook, or twitter, etc, you should expect your content to be packaged, sold and used - the ire placed at MacRumors seems misplaced at best.

I suppose if anyone one person who doesn't want their content to be used in ways they don't approve of, then they need to reconsider how they interact with social media, and any platform that uses the data, i.e., google.

I've largely stopped using many social platforms, but I'm under no illusions that my interactions here and elsewhere are not used to the profit of others.