I don't like what "AI" is doing for original content

[StereoNET]

Allow me to rant, please.

 

Back in November a new brand launched. As per usual processes, the manufacturer sent out press releases - it contained the basic info journalists needs to construct a story, along with images, tech specs etc. As is often the case, an embargo date was set, meaning the information cannot be shared or published until that time. StereoNET has always honoured embargoes, but unfortunately with the rise of blogs, youtubers, and influencers without any editorial background, training, or experience, these days it is becoming increasingly common for these to be totally ignored (or misunderstood).

With web publishing, it's very much a case of the "the early bird catches the worm". At least, to get the traffic and Google indexing, that is very much important.

 

The manufacturer reached out to StereoNET prior to the embargo to offer a world exclusive review. In other words, of all the publications, they gave us the opportunity to review the very first product and have access to the only review sample that had been made available anywhere. We have no commercial obligations to this brand, but liked the story behind it and appreciated the opportunity, so we agreed. No money changed hands, it was purely a good news story that we wanted to support.

 

The embargo lifts and the news breaks around the launch of this brand. In addition to the news, we also published our exclusive review. 

 

So you can imagine my surprise when another review turns up at the same time of the same product at another publication. We immediately reached out to our contacts and were assured that no sample was provided to them, and were told it would have been practically impossible for them to have had access to a sample to conduct a review. It was concluded by all, that the review was entirely fabricated. 

 

Roll forward a month. A little late to the party, I discovered Lucca Chesky, son of famed musician and label owner David Chesky, had released his own loudspeaker. So on 18th December I wrote a short piece based on information predominantly from the product listing on their website and published that here: https://stereonet.com/news/cheskys-next-generation-delivers-new-loudspeaker

 

No press release was ever received on this one, nor do I believe one was distributed at any point. I often add comment based on industry knowledge or experience into these news pieces where relevant. I did so in this case with these opening paragraphs:

 

Quote

David Chesky needs no introduction if you've been around high-fidelity audio in the last few decades. His label, Chesky Records, is a staple in most audiophiles' vinyl or CD collections.

 

But follow any sportsperson, musician, movie star, or, in this case, a renowned musician and record label head long enough, and quite often, you'll see their offspring emerge following in their footsteps.

 

Then I chanced upon another article on a site I am not familiar with, Stereo Index, on the 23rd December, titled "Chesky Audio Unveils the LC1 Speaker". https://www.stereoindex.com/speakers/chesky-audio-lc1-the-first-loudspeaker-from-new-manufacturer

 

What caught my particular attention was this particular paragraph:

 

Quote

David Chesky is a well-known figure in the realm of high-fidelity audio. His label, Chesky Records, is a prominent presence in the collections of numerous audiophiles, whether in vinyl or CD format. It is common to observe that the children of prominent athletes, musicians, or movie stars often pursue similar paths. However, this is not always the case.

 

It is not unusual to use AI tools to assist in publishing today. We often use them to help generate a particular format of image we need to support an article, or working within Grammarly, to name one example, to speed up the editing and sub-editing processes for news type articles. When I wrote the above article, Grammarly AI wanted to rewrite my paragraph above to exactly as shown in bold above on the article at Stereo Index. Now AI is not 100% accurate, often does not understand context, and more often that not I don't agree with most of its suggestions. As was the case here, I rejected that exact suggestion as it changed the meaning of the sentence entirely as you would agree when you read the article in complete context. 

 

But it confirmed without doubt, that my article was the source of this original content. The author had simply lifted the text, dropped it into an AI tool for content generation and proceeded to publish that story. There are other tell-tale signs through the piece too. 

 

What can be done? Likely nothing at all. While obvious to me, it would still be very hard to prove, and in the world of copyright law, I expect that the wording has been changed enough to get around dated laws regarding this.

 

But given these two examples in just the last month, and more we have witnessed over the last year, I suspect we're only going to see more and more entirely fabricated content. This is highly concerning, moreso in the case of product reviews which play a vital role for consumers seeking qualified opinions before making purchase decisions. News (at least, in our industry), is somewhat less critical or important, I guess. I mean, the world's news outlets are basically regurgitating the same press information from the same news wire services or Reuters around the world and have been for decades already. But when it comes to original content, or "Reviews" I do find this alarming and concerning, particularly with how easy it is now to generate content from just a few inputs.

 

I guess my point is, we've always known that, "just because it's on the internet doesn't mean it's true" has never been more relevant than it is now. At least here at StereoNET, we're a little more old school and still believe in writing original content. Long term, eventually we'll of course be replaced by millions of AI-generated blogs and sites, and our model will no longer be commercial sustainable. We don't subscribe to the affiliate link-driven editorial model now employed by What Hi-Fi? and a growing number of others, relying on support instead through genuinely invested brands and companies similarly invested in our industry itself. I know this is not the future, and some brands would now rather send a free product to an influencer with 150K subscribers who often knows little about the hobby or product in return for a review or "plug", instead of supporting independent publications, employing experts producing original content like StereoNET and many others. I don't hold that against the brands themselves. They have commercial goals themselves they need to achieve to justify their existence and, let's face it, the influencer or "ambassador" model works well too. 

 

My own little experience so far with this, though, gives me cause for genuine concern about the future of publishing, news, and information. 

[Steff]

Plagiarism is just another byproduct (intended) of these services.

 

I know I'm stemming myself against the tide here, but the sooner people stop calling it "AI" and naming it what it really is: machine learning, Large Language Learning Models, the clearer expectations can become. There's no magic or intelligence in these models (yet), they're no more than parrots with a large (context) vocabulary.

 

To call it AI just reproduces the aggressive billionaire and big corporate tech marketing agenda.

[mrbuzzardstubble]

4 minutes ago, Steff said:

Plagiarism is just another byproduct (intended) of these services.

 

I know I'm stemming myself against the tide here, but the sooner people stop calling it "AI" and naming it what it really is: machine learning, Large Language Learning Models, the clearer expectations can become. There's no magic or intelligence in these models (yet), they're no more than parrots with a large (context) vocabulary.

 

To call it AI just reproduces the aggressive billionaire and big corporate tech marketing agenda.

I believe one day plagiarism will be a thing of the past. You may call me a dreamer, but I'm not the only one...

[sir sanders zingmore]

In the words of the late, great Terry Pratchett

”Real stupidity beats artificial intelligence every time.”

 

Seriously though, this is a real problem and I fear it’s only going to get worse and harder to spot as these models get better

[buddyev]

The great early dream of the internet was a democratisation of information and sidestepping the gatekeepers  - yeh right.
I think we all understand now that going to authoritative and trusted sources which edit and curate information is the only counter to the slew of rubbish. It could well be that rather than drowning in AI garbage the sites which offer this will become ever more important. 

[Batty]

George Orwell was right

[ZEN MISTER]

So-called AI has evolved as a duality: a definite advance for aspects of medicine, environment concerns, and other venues of progressive human evolution .

The other side of the coin is its availability of parasitic tendencies . Within the field of creative endeavour, it has no worthy presence of note. 

In my own experience, I have contributed numerous articles to SNA. I was pleased to see articles. Interviews and reviews taken up on the artist, producer or musicians own sites.

Google my David Chesky interview, and it is right up there in the 'hits' on google .

An article on The Electric Flag continues to clock up reads years after original publish date  .

But I would be very pissed to find anything I wrote ripped off, altered, or otherwise vitiated .

AI should have had ' water marking ' ability from early in its development .

Give mankind a tool of potential good, and mankind will find a means to turn it into a regressive tool , both subtle and bludgeoning.

Billy ( the real one).

 

 

Edited December 24, 2024 by ZEN MISTER

[audiofeline]

I find it quite alarming how AI has influenced so many in such little time.  People speak of AI's "creativity", but all it can do is to look at what's been done before and create mashup's of them.  Creative in a sense, but not truly original.  So AI will provide more of an average of past creativity, rather than extend our creative worlds. 

 

But it also has other effects.  Because AI makes it so easy to "write" something informative, people are going to lose the ability (or if young, not acquire the ability) to analyse information critically and present a cohesive argument.  The worries of schools and Universities of plagiarism are eclipsed by the AI frontier. 

 

Marc, I admire the integrity you bring to your writing and to the SNA world.  Very unfair that your hard work is being appropriated by others.  In reference to your above comment, I think you have a copyright claim against that other publication - they have infringed your original content, and you can demonstrate it.  The fact that they did it via AI does not relinquish their decision to publish without checking the validity of the content.  It's your words, and they substantially stole them.  But a copyright claim will be a costly path to go down.  You might be right, but the personal cost to you may over-ride any rewards gained.  Life can be unfair.  At the very least, I think a letter of complaint to the editor who published your work is warranted. 

 

 

 

 

[bob_m_54]

On 24/12/2024 at 11:26 AM, Batty said:

George Orwell was right

What?

“Four legs good, two legs bad.”

[Steff]

ps. for anyone with interest and a little time...https://www.elementsofai.com/