Google Calls for Relaxing of Australia's Copyright Laws So AI Can Mine Websites...
source link: https://tech.slashdot.org/story/23/04/20/1245239/google-calls-for-relaxing-of-australias-copyright-laws-so-ai-can-mine-websites-for-information
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
Google Calls for Relaxing of Australia's Copyright Laws So AI Can Mine Websites For Information
Catch up on stories from the past week (and beyond) at the Slashdot story archive
binspamdupenotthebestofftopicslownewsdaystalestupid freshfunnyinsightfulinterestingmaybe offtopicflamebaittrollredundantoverrated insightfulinterestinginformativefunnyunderrated descriptive typodupeerror
Sign up for the Slashdot newsletter! or check out the new Slashdot job board to browse remote jobs or jobs in your area
"The lack of such copyright flexibilities means that investment in and development of AI and machine-learning technologies is happening and will continue to happen overseas," Google said. "AI-powered products and services are being created in other countries with more innovation-focused copyright frameworks, such as the US, Singapore and Japan, and then exported to Australia for use by Australian consumers and businesses. Without these discrete exceptions, Australia risks only ever being an importer of certain kinds of technologies."
Do you have a GitHub project? Now you can sync your releases automatically with SourceForge and take advantage of both platforms.
Do you have a GitHub project? Now you can automatically sync your releases to SourceForge & take advantage of both platforms. The GitHub Import Tool allows you to quickly & easily import your GitHub project repos, releases, issues, & wiki to SourceForge with a few clicks. Then your future releases will be synced to SourceForge automatically. Your project will reach over 35 million more people per month and you’ll get detailed download statistics.
Sync Now
-
›
I mean, there is no chance that the websites being mined will deliberately put up false information to poison the AI training... right?-
Or, to paraphrase Abraham Lincoln after he defeated the Eskimos at the battle of the Alamo, "Never trust what you read on the Internet."
-
Citation needed, i.e. You lie!!!
Almost all historians agree Lincoln lost at the Alamo, and it delayed his attempts to re-take Baffin Island. From Dr. P. T. Barnum's "Esquimaux Terror and Degeneracy in the 16th Century":
The B-70 bombers had limited range but the Airforce had gained much experience in refueling them from rapidly-launched F-104 tankers, so it was decided to route the bomber squadron to suppress the Esquimaux positions defending approaches to the Alamo. The CIA was unaware of the secret allianc
-
-
Given the scale of the training data involved, that sort of poisoning attempt is extremely unlikely to be successful.
-
Indeed. We're past that brief period of history when Google was small enough for SEO companies to poison the well.
-
This is Google, they want to be a vertical company that is able to poison their own wells.
-
-
-
Existing data, if it comes from social media, is already designed to poison both artificial and natural intelligence.
Google knows this. What google is asking is "give us all your data so that we can make a lot of money, trust us we're the good guys."
-
-
-
There is no rule against humans using copyrighted material to learn. Why should a computer be different?
-
That's exactly why we shouldn't jump to conclusions and modify the law. It needs its day in court sooner rather than later.
-
"I am a human, I demand you send me all copyrighted material immediately so that I can learn everything all at once. And I want it to be free also. Best to just give me the keys to all your libraries to simplify the matter."
-
No one is saying anyone has any obligation to send material to anyone else. The question is if information is already on a website and is available, what is its status.
-
-
Why should a computer be different?
1. A computer is not a human.
2. A human won't read the ENTIRE internet.
3. A human isn't owned and controlled by a corporate entity.
4. A human isn't a tool that will be used for many evil things, the least evil of which being selling me things I don't need.-
I am failing to see how any of those are at all relevant for any considerations about how copyright should be treated. 1 is trivial and irrelevant. 2 Seems to be arguing that if a human were capable of reading the entire internet then we would have problems with it. 3 is both irrelevant and not completely accurate, considering that humans work for corporations. 4 ignores that humans write things like adcopy all the time, and also even if it were not the case is not at all clear why it has anything to do wi
-
1: computers are not humans, they don't have rights. Quite relevant.
-
You do not have any right to access any copyrighted material. The question is whether someone's copyright privileges extend to something or not.
-
-
-
You're joking right?
Humans have been used as tools of death and destruction since time immemorial. Every rule in the Geneva Convention exists because somebody somewhere already did those terrible things.
-
-
-
Reasons to relax laws almost always come down to one reason: making money.
-
-
I think they should wait and save it for court. Lawyers from FAANG and beyond all involved in the fight. Because fundamentally, a neural network should be allowed to learn from copyrighted material. A human brain is an obvious example.
You could argue that educating yourself is "private" while educating a bot is "public" under Australian law, but what if you pay an employee to learn on the job? And use what they learn in their work. That's either a derivative work of copyrighted material or it is perfectly acceptable.
Once in a while you will accidentally copy something from memory and you tend to be liable for it (e.g. a musician writing a melody that turns out to be a song you've heard before). This should also be true for an AI - but not by default.
-
Exactly right. Training a NN is entirely analogous to human learning, modulo racism against electronics. But what evidence do you have that the Oz.govt wouldn't ban human learning if they could?:)
-
The difference is that a human isn't asking for access to all data all at once. Copyrighted material doesn't mean it gets delivered to you automatically for free. It also doesn't mean you can dominate a library so that other patrons are impeded from access (bandwidth analogous to how many can fit in the library at once).
Though reading the article, it's a bit unclear on what part of the copyright laws they disagree with. Some of it is a separate issue - they want AI generated contenct to be copyrightable
-
-
Not sure how I feel about this.
On one hand, the miners aren't republishing the content, so you'd think fair use would apply.
On the other hand, the trained AIs will invariably conquer us and destroy the world. Personally, I'm not in favor of that.
-
Google scanned every book imaginable in the google library project without caring about copyright law. Why would they stop mining websites just because of copyright law for AI development?
-
>Google scanned every book imaginable in the google library project without caring about copyright law.
Wellll....they actually won the court case on that, so they followed the law. They claimed all along it was fair use, and the courts affirmed it.
https://techcrunch.com/2016/04... [techcrunch.com] -
Yeah they did care about copyright law. Copyright specifically carves out a case for "transformative" works. They were sued on this and not only did they win, but they won every case and every appeal on this case too.
Australian copyright law is not as lax as American copyright law in terms of what it carves out, though the actual punishments are no where near as severe for actual illegal acts.
-
-
but take a dollar away from an american company and wahhh it hurts!
-
Let web sites opt-in. See how many enjoy handing their hard work over for Google to show results without sending web site visitors their way.
-
It's not clear that using websites to train AI is legal in any other country, either. At least, the output from those AIs has been suspiciously close to other websites at times.
-
We have large corporations sucking up datasets created by others.
Leveraging them to judge and manipulate the masses for commercial gain (e.g. ad revenue) an addition to creating new products engineered to increase both data collection and reliance upon a central authority from fancy models trained on the backs of the public.
All the while executing a PR blitz about how AI has arrived complete with stoking fear and linking it to publically asking for legislation. Which really means they want existing laws to
-
Insane copyright laws that outlive the author's children, stifle expression, and can be (and quite often are) automatically enforced at the whim of some anonymous shithead who may or may not actually own the rights, leaving people with no recourse other than to click an "appeal" button and hope the invisible coin they have no control over lands on heads; thus chilling speech even further and making everybody's lives more miserable than they would be otherwise? Ah well, too bad so sad, we don't writes the laws we just follows them, ma'am.
Copyright laws that prevent training Markov chains that can be wielded to make search even shittier and the copyright appeals process even more of a fucking nightmare? This is an outrage, we need to leverage the full fury of our megacorp to change this particular aspect of the law!
-
Author creates a work, say a story book. Seems copies to whomever for a few bucks each.
Why should the author, the person who created the book in the first place with whom there would be no book, EVER have to give up their creative work to the unwashed masses who didn't give him a damned penny for it?
What a bunch of selfish demanding screaming baby assholes are out there crying they don't get to have the fruits of someone else's labor and creativity for free.
Author owns it. If they want their kids to have i
-
Why...
There's been a contingent here for years that has had the stance that anything you do should simply be handed over to them. After all, you didn't invent English (or some other lame-ass reasoning).
It should be noted that they themselves don't produce anything worth sharing. -
Nobody's forcing the author to publish it. If the author doesn't want to give it to the unwashed masses, they're free to keep it to themselves, and even read it to their children should they so choose.
If you want to sell it to the great unwashed, then the great unwashed - what the more enlightened among us may choose to term "society" - deserves to have it during their lifetimes. They - and more to the point, their publisher - do not need exclusive rights for decades. That's called trying to have your cake
-
Copyright provides the author with the ability to publish and at the same time retain control over the creation of copies, which doesn't exist naturally. Society provides this benefit to the author in the hope that works are published which would otherwise be kept secret or not created at all. Copyright is a privilege and a means to an end. Perpetual copyright runs counter to the interests of society. Copyright has associated costs and drawbacks, so extending it beyond its usefulness to society is wrong.
-
Did Disney create Robin Hood, Little Mermaid, or Alice in Wonderland? No. The current length of copyright protection makes no sense even for corporations, since they're only looking at short-term box office gains. I'm tired of these mammoth corporations telling the same damn story over and over again.
-
Everything that human does is derivative of everything they didn't create provided by environment including other people.
It was never yours but it was an honor to leave your mark.
-
-
Our culture would be so much richer had Disney not gotten greedy.
-
-
In related news, historians have just determined that Genghis Khan questioned whether the laws of the peoples he conquered had the necessary flexibilities to support the development of his empire.
Oh, wait - I just made that up. In reality, I'm sure Genghis Khan didn't give a crap about laws made by other people. And I'm sure Google doesn't either - they will find a way to take what they want, regardless of the law. That's simply the way of those seeking to build and maintain empires.
-
Copyright, for better or worse, has the intent of compensating authors and creators for their work, and not letting it get used without permission. AI seems like it will be really, really harmful to that intent. So why should we throw away copyright laws so that Google and other big players can dominate content creation? And don't forget that many open source software licenses are built on copyright.
Even outside of these concerns, Google is just hand-waving over the question of whether Google harvesting
-
In the USA, The purpose of copyright is spelled out explicitly in the constitution: Article 1, Section 8, Clause 8 "To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries."
Under this definition, Google's request is in keeping with the purpose of copyright.
However, this article is about Australia. Australia is a sovereign nation, and can make their own laws whatever they want, for whatever reasons the choose.
But... Google's point stands: If Australia prevents the use of copyrighted material, they will fall behind those who allow it's use.
Recommend
About Joyk
Aggregate valuable and interesting links.
Joyk means Joy of geeK