Urban75 Home About Offline BrixtonBuzz Contact

Artificial Intelligence Developments (ChatGPT etc)

EU rules on privacy and social media suggest you can regulate big tech. AI enforcement seems very hard to do, but will be easier to monitor the big tech companies than it will open source and bad faith actors.
Those big tech companeis will want advantages to counterbalance all of that. They'll be banging the drum of "deepfake misinformation danger" to try and keep independent AIs out of the hands of the rest of us.

That's my main concern, that there will be a narrative to keep the AIs under the control of large entities only because the rest of us can't be trusted, that some AI dangers won't be tolerated while others are.
 
I have done an experiment. I created a webpage largely from ChatGPT outputs and am watching to see what G makes of it. So far it has dropped with hardly a trace which isn't good. G says it can recognise AI generated texts and won't allow them to do well in rankings. I had partially rewritten the page in my own writings but it seems I didn't do that enough. That or the construction of the page is just not suited to rank at all. More tests to come as I try to understand what is going on.
 
Why Nature will not allow the use of generative AI in images and video

Saying ‘no’ to this kind of visual content is a question of research integrity, consent, privacy and intellectual-property protection.
[...]
Artists, filmmakers, illustrators and photographers whom we commission and work with will be asked to confirm that none of the work they submit has been generated or augmented using generative AI

Also Spaceweather has had this statement at the top of the page for a while:

Text created by ChatGPT and other Large Language Models is spreading rapidly across the Internet. It's well-written, artificial, frequently inaccurate. If you find a mistake on Spaceweather.com, rest assured it was made by a real human being. This is an AI Free Zone!
 
I have done an experiment. I created a webpage largely from ChatGPT outputs and am watching to see what G makes of it. So far it has dropped with hardly a trace which isn't good. G says it can recognise AI generated texts and won't allow them to do well in rankings. I had partially rewritten the page in my own writings but it seems I didn't do that enough. That or the construction of the page is just not suited to rank at all. More tests to come as I try to understand what is going on.

Did you ask ChatGPT to write the content using the keywords you want to rank in google for?

Example :

Please write a 500 word article about investing. The article needs to rank well in google for the following terms "Gold sovereigns", "Pensions" and "Retirement"
 
Did you ask ChatGPT to write the content using the keywords you want to rank in google for?
No, I didn't want to be obviously trying to rank well

I asked ChatGPT a few relevant questions on a specific topic and then massaged the responses into an article about that topic. I rewrote some of it and left other parts as ChatGPT had output.

Then I uploaded it and a few days later confirmed that G had crawled and indexed the page but it didn't rank for the subject the page was about. It might be early days, I will check again soon but if it completely tanks in G I might have to rethink how I use ChatGPT.
 
No, I didn't want to be obviously trying to rank well

I asked ChatGPT a few relevant questions on a specific topic and then massaged the responses into an article about that topic. I rewrote some of it and left other parts as ChatGPT had output.

Then I uploaded it and a few days later confirmed that G had crawled and indexed the page but it didn't rank for the subject the page was about. It might be early days, I will check again soon but if it completely tanks in G I might have to rethink how I use ChatGPT.
[Plugin] 💖 SEO.app is the first SEO optimization assistant for ChatGPT!
 
Lawyers submitted bogus case law created by ChatGPT. A judge fined them $5,000 - AP News

At a hearing earlier this month, Schwartz said he used the artificial intelligence-powered chatbot to help him find legal precedents supporting a client’s case against the Colombian airline Avianca for an injury incurred on a 2019 flight. (...)

The judge said the lawyers and their firm, Levidow, Levidow & Oberman, P.C., “abandoned their responsibilities when they submitted non-existent judicial opinions with fake quotes and citations created by the artificial intelligence tool ChatGPT, then continued to stand by the fake opinions after judicial orders called their existence into question.” (...)

In a separate written opinion, the judge tossed out the underlying aviation claim, saying the statute of limitations had expired.

🤣
 
The people paid to train AI are outsourcing their work… to AI :thumbs:

Many companies pay gig workers [...] to complete tasks that are typically hard to automate, such as [...] labeling data [which] is then fed into AI models to train them. The workers are poorly paid and are often expected to complete lots of tasks very quickly.
[...]
a team of researchers [...] hired 44 people on the gig work platform Amazon Mechanical Turk to summarize 16 extracts from medical research papers. Then they analyzed their responses using an AI model they’d trained themselves that looks for telltale signals of ChatGPT output [...] They estimated that somewhere between 33% and 46% of the workers had used AI models like OpenAI’s ChatGPT.
[...]
Using AI-generated data to train AI could introduce further errors into already error-prone models.
 
I asked ChatGPT about primes ending in 2 :facepalm:

It gave me a list of numbers starting with 2, 3, 5, 11, 31, 41...
[...]
it then claimed that out of those 50 are odd while the others are even, and that therefore 50% would be odd.
[...]
I asked ChatGPT if there are only four prime numbers that end in 2. It confirmed this saying the four prime numbers ending in 2 were 2, 3, 5, 7
[...]
I then asked ChatGPT if there are five prime numbers that end in 2. It apologized for its oversight and confirmed this saying the five prime numbers ending in 2 were 2, 3, 5, 7, 11
 
If you tell it it's wrong it often apologises and then repeats the error. I suppose it gets it wrong because it doesn't know the answer and so isn't able to self-correct (and won't take corrections from users) but it also seems unable to reply that it doesn't know or isn't sure.
 
this looks great, open source and free image generator as good as if not better than all the competition
i get the feeling theres going to be a lot of good open source AI stuff out there


SDXL 0.9 is the latest version of the Stable Diffusion text-to-image suite of models developed by Stability AI, a London-based artificial intelligence startup. It was announced on June 22, 2023 and claims to produce more realistic and detailed images than its predecessor, SDXL beta. It has a parameter count of 6.6 billion and uses two CLIP models to process text prompts and generate images at a resolution of 1024x1024. It also offers functionalities such as image-to-image prompting, inpainting, and outpainting. The model is currently available through ClipDrop, with an API release coming soon, and an open release planned for mid-July▼ Link(s) From
.✩ TRY SDXL 0.9 for FREE Here: https://www.futurepedia.io/tool/clipdrop
 
I see that GPT is as bad at writing music in the style of Bach as it is at writing essays. It’s missing everything that makes Bach great, which includes the way he constantly modulates through keys in imaginative and unexpected ways. Through its counterpoint but also this modulation an emotional intensity is produced that he keeps hanging and hanging until the final desperate relief. What’s produced by the AI is nothing but basic baroque as I would expect to be taught to GCSE students. It’s boring and is as to Bach as a stick figure is to Vermeer.
 
Interesting I thought.

There is this guy who is an online SEO expert who also has his own Ad agency. I saw him last week saying that ChatGPT makes your life easier making content so he recommends people use it.

I thought ChatGPT pages were supposed to tank in Google SERPS and quite a few people seem to have made experiments that have evidenced this.

Anyhow this SEO guy has been proven to know his stuff so I will investigate further.
 
Authors file a lawsuit against OpenAI for unlawfully ‘ingesting’ their books

Two authors have filed a lawsuit against OpenAI, the company behind the artificial intelligence tool ChatGPT, claiming that the organisation breached copyright law by “training” its model on novels without the permission of authors. [...] Awad and Tremblay believe their books, which are copyrighted, were unlawfully “ingested” and “used to train” ChatGPT because the chatbot generated “very accurate summaries” of the novels
[...]
The complaint said that OpenAI “unfairly” profits from “stolen writing and ideas” and calls for monetary damages on behalf of all US-based authors whose works were allegedly used to train ChatGPT. Though authors with copyrighted works have “great legal protection”, said Saveri and Butterick, they are confronting companies “like OpenAI who behave as if these laws don’t apply to them”.
 
This issue has been around for a while, also with the likes of Google. Google has the added cheek of maintaining a cache of people's web pages so that searchers can even see actual web pages without leaving Google's property and without even visiting the publishers actual web site.

No one was prepared to rock the boat with G about this violation which set a precedent for the likes of OpenAI.
 
Back
Top Bottom