Perplexity AI decries News Corp's 'simply false' data scraping claims
- Reference: 1729886408
- News link: https://www.theregister.co.uk/2024/10/25/perplexity_news_corp_data/
- Source link:
At the start of the week, News Corp took [1]legal action on behalf of its publications, claiming in court [2]documents [PDF] that Perplexity was posting whole chunks of the newspaper's articles, wrongly attributing facts to the sources, and that Perplexity ignored them when the issue was raised. News Corp wants $150,000 for every proven infringement.
But the News Corp filing contains "falsehoods," according to Perplexity.
[3]
"The suggestion that we never responded to outreach from News Corp is simply false: they reached out; we responded the very same day; instead of continuing the dialogue, they filed this lawsuit," Perplexity said in a sharply worded [4]blog post .
[5]
[6]
Secondly, the startup denies claims that Perplexity's engine merely regurgitates text. The team asserts that where excerpts from articles can be linked, the AI engine provides those links.
"We have learned in the short time since this lawsuit was filed, a disturbing trend in these types of cases: The companies that are suing make all kinds of salacious allegations in their complaints about all kinds of seemingly bad things they were able to coax the AI tools to do—and then, when pressed in the litigation for details of things like how they achieved such obviously unrepresentative results, they immediately disavow the very examples they put in the public record, and swear they won’t actually use them in the case," it said.
[7]
"We presume that is what will happen here. And that will tell you everything you need to know about the strength of their case."
[8]Major publishers sue Perplexity AI for scraping without paying
[9]Everyone wants better web search – is Perplexity's AI the answer?
[10]Cloudflare debuts one-click nuke of web-scraping AI
[11]LinkedIn started harvesting people's posts for training AI without asking for opt-in
The Perplexity crew points out that it already has content-sharing deals in place with Time, Fortune, and Der Spiegel, and would have been happy to work with News Corp on a similar deal. Rather cheekily, they also point out that the Wall Street Journal, another News Corp publication, voted them the best AI chatbot earlier this year.
"There are around three dozen lawsuits by media companies against generative AI tools," the AI startup notes.
"The common theme betrayed by those complaints collectively is that they wish this technology didn’t exist. They prefer to live in a world where publicly reported facts are owned by corporations, and no one can do anything with those publicly reported facts without paying a toll."
News Corp declined to comment on Perplexity's claims. ®
Get our [12]Tech Resources
[1] https://www.theregister.com/2024/10/22/publishers_sue_perplexity_ai/
[2] https://regmedia.co.uk/2024/10/21/perplexity.pdf
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZxwVBu8-7pcEO11KTVXwfgAAAJY&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[4] https://www.perplexity.ai/hub/blog/about-the-dow-jones-lawsuit
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZxwVBu8-7pcEO11KTVXwfgAAAJY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZxwVBu8-7pcEO11KTVXwfgAAAJY&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[7] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZxwVBu8-7pcEO11KTVXwfgAAAJY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[8] https://www.theregister.com/2024/10/22/publishers_sue_perplexity_ai/
[9] https://www.theregister.com/2024/01/05/perprexity_ai_search_engine/
[10] https://www.theregister.com/2024/07/03/cloudflare_ai_blocks/
[11] https://www.theregister.com/2024/09/19/linkedin_ai_data_access/
[12] https://whitepapers.theregister.com/
WTF
"where publicly reported facts are owned by corporations'"
Well, it's the corporations that made those facts public for you to lift.
Next argument.
"Publicly visible art is owned by the artists", so we'll steal that content as well.
Now for a shower after defending News Corp.
In building an AI, one of the biggest concerns has to be the cleanliness of the initial dataset. I’d argue that this alone makes scraping the broader internet a very foolish thing to do if the object is to build a tool which is useful and expert in its field. That said, at least the internet as a whole has a broad range of views that one can draw on - so at least it will be equally awful on a wide range of views and opinions.
Training any dataset on a Murdoch organ like the New York Post though? I mean, fine if you want to build a uniquely fascist AI with objectionably racist views. Like building an AI out of the Daily Fail or the Maily Telegraph. I wouldn’t even want to build an AI out of the Guardian - although I suspect in that case the worst thing that could be said about it is that it had a perplexing predilection for home made muesli.
news Corp are the scum (via the news of the world) who left fake voicemails on a murdered teenagers voicemail to trick the parents into thinking she was still alive.
So they could basically run shitty evil monstrous front page "she called home" type stories.
Everyone at news corp should be banned from pretending to be a journalist for life.
If I rewrote this and every other Reg article with mild paraphrasing along the lines of:
"AI newcomers Perplexity have responded to a legal claim that they're stealing content from..."
and then based my business on publishing that content, I'd (rightly) get sued.
Mind you they should be sued just for the sickening tone of that blog post alone