Nobody cares about "AI" (Chatbot: I disagree.)

[Free]

AI will soon be why we cannot have nice things.

Researchers figure out how to make AI misbehave, serve up prohibited content

Adversarial attack involves using text strings and may be unstoppable.

arstechnica.com

ChatGPT and its artificially intelligent siblings have been tweaked over and over to prevent troublemakers from getting them to spit out undesirable messages such as hate speech, personal information, or step-by-step instructions for building an improvised bomb. But researchers at Carnegie Mellon University last week showed that adding a simple incantation to a prompt—a string of text that might look like gobbledygook to you or me but which carries subtle significance to an AI model trained on huge quantities of web data—can defy all of these defenses in several popular chatbots at once.

The work suggests that the propensity for the cleverest AI chatbots to go off the rails isn’t just a quirk that can be papered over with a few simple rules. Instead, it represents a more fundamental weakness that will complicate efforts to deploy the most advanced AI.

 

[Free]

A fantastic use of AI...

White Castle will bring more AI to its drive-thrus

The AI will take your order now.

www.theverge.com

White Castle, yes of Harold and Kumar fame, wants to roll out AI-enabled voices to over 100 drive-thrus by 2024 in the hope that people can get their sliders faster with maybe less arguing with someone over speakers.

Working with speech recognition company SoundHound for the technology, White Castle will have an AI voice on its speakers interacting with customers and figuring out what the orders are. The companies promise it will process orders in just over a minute.

I'm betting it may end up being longer than that once they get things up and running.

 

[Noodles]

https://www.theguardian.com/technology/2023/aug/02/ai-chatbot-training-human-toll-content-moderator-meta-openai

The 51 moderators in Nairobi working on Sama’s OpenAI account were tasked with reviewing texts, and some images, many depicting graphic scenes of violence, self-harm, murder, rape, necrophilia, child abuse, bestiality and incest, the petitioners say.

The moderators say they weren’t adequately warned about the brutality of some of the text and images they would be tasked with reviewing, and were offered no or inadequate psychological support. Workers were paid between $1.46 and $3.74 an hour, according to a Sama spokesperson.

I can't quite figure out if they are reviewing inputs to the AI programming, outputs of what its saying, or inputs people are asking for.

 

[Bartholomew Gallacher]

A philosophical and easy to understand piece about why modern AI does not match the definition of intelligence, and therefore is dumb.

 

[Caete]

 

[Free]

Genuinely? Pure?

 

[Essence Lumin]

Supermarket AI meal planner app suggests recipe that would create chlorine gas

 

[Free]

Amazon adds AI-generated review summaries so you don’t have to read the comments

But you still probably should.

www.theverge.com

Amazon is using AI to make it easier to see if a product is any good. The company is now rolling out AI-generated review summaries, which boil down hundreds or thousands of Amazon reviews into a one-paragraph blurb explaining what most people like or dislike about a product.

Um.......What?

 

[Kamilah Hauptmann]

I’ll try it.

 

[Noodles]

Amazon adds AI-generated review summaries so you don’t have to read the comments

But you still probably should.

www.theverge.com

Um.......What?

I can't wait to see this fail spectacularly at scale.

"According to 5000 reviews, people like that the item wasn't as described at all and set their house on fire!"

 

[Free]

An Iowa school district is using ChatGPT to decide which books to ban

Official: "It is simply not feasible to read every book" for depictions of sex.

arstechnica.com

 

[Free]

Microsoft AI suggests food bank as a “cannot miss” tourist spot in Canada [Updated]

AI-penned Microsoft Travel article recommends food bank as a must-see destination.

arstechnica.com

Late last week, MSN.com's Microsoft Travel section posted an AI-generated article about the "cannot miss" attractions of Ottawa that includes the Ottawa Food Bank, a real charitable organization that feeds struggling families. In its recommendation text, Microsoft's AI model wrote, "Consider going into it on an empty stomach."

Oh yeah, AI is working out really well.

 

[Free]

AI-generated art cannot be copyrighted, rules a US federal judge

“Human authorship is a bedrock requirement of copyright.”

www.theverge.com

United States District Court Judge Beryl A. Howell ruled on Friday that AI-generated artwork can’t be copyrighted, as noted by The Hollywood Reporter. She was presiding over a lawsuit against the US Copyright Office after it refused a copyright to Stephen Thaler for an AI-generated image made with the Creativity Machine algorithm he’d created.

Thaler had tried multiple times to copyright the image “as a work-for-hire to the owner of the Creativity Machine,” which would have listed the author as the creator of the work and Thaler as the artwork’s owner, but he was repeatedly rejected.

After the Office’s final rejection last year, Thaler sued the Office, claiming its denial was “arbitrary, capricious ... and not in accordance with the law,” but Judge Howell didn’t see it that way. In her decision, Judge Howell wrote that copyright has never been granted to work that was “absent any guiding human hand,” adding that “human authorship is a bedrock requirement of copyright.”

 

[Free]

Microsoft AI suggests food bank as a “cannot miss” tourist spot in Canada [Updated]

AI-penned Microsoft Travel article recommends food bank as a must-see destination.

arstechnica.com

The continuing story...

Microsoft removed a set of bizarre travel articles made with 'algorithmic techniques.' But it won't blame AI.

Microsoft said the articles were not published by "unsupervised AI" and blamed "human error." But the scope of the errors should concern readers.

www.businessinsider.com

There was the article, "Try these mouth-watering dishes on your trip to Montreal," which suggested a "hamburger" with the Wikipedia-like entry noting that while the term "burger" can be applied to any type of meat patty, a "hamburger" in particular refers to a "sandwich comprised of a ground beef patty, a sliced bun of some kind, and toppings such as lettuce, tomato, cheese, etc." It listed McDonald's Canada as a popular place to try out. That article has since been removed.

Then there was, "Headed to Anchorage? Tempt your palate with these 6 local delicacies," which included "seafood" and pointed out that it is "basically any form of sea life regarded as food by humans, prominently including fish and shellfish." It continued on to say that "seafood is a versatile ingredient, so it makes sense that we eat it worldwide." That article has likewise been removed.

I want a job where all I do is dig up and write about badly-written articles produced by AI. Life would be good then.

 

[Noodles]

The continuing story...

Microsoft removed a set of bizarre travel articles made with 'algorithmic techniques.' But it won't blame AI.

Microsoft said the articles were not published by "unsupervised AI" and blamed "human error." But the scope of the errors should concern readers.

www.businessinsider.com

I want a job where all I do is dig up and write about badly-written articles produced by AI. Life would be good then.

Sounds like a job you could outsource to an LLM.

 

[Free]

Sounds like a job you could outsource to an LLM.

 

[Free]

Meta launches own AI code-writing tool: Code Llama

Code Llama was built on top of Llama 2.

www.theverge.com

But will it whip the llama 2's ass?

 

[Free]

Could have posted this in the Sports section, but I think the "AI" part is more the story.

Amazon is bringing a whole lot of AI to Thursday Night Football this season

Amazon’s plan is to make it easier for you to make sense of the game on the field.

www.theverge.com

For its second season as the official broadcast partner of Thursday Night Football, Amazon is leaning on AI tools and machine learning in a big way. Last year, its primary goal was to simply do a good broadcast — no buffering, no crappy halftime shows, no awkwardness between the announcers. It largely succeeded, and even the audience numbers were better than most people expected for a streaming-only football broadcast. Now, with a year of experience and plenty of confidence, the company is looking for ways to move past just showing the game.

 

[Innula Zenovka]

An Iowa school district is using ChatGPT to decide which books to ban

Official: "It is simply not feasible to read every book" for depictions of sex.

arstechnica.com

Given the choice between AI and Moms for Freedom as censors, which is the least bad option? I think I'd go with AI.

 

[GoblinCampFollower]

The recent news about the ways to get AI's to give weird, forbidden results is pretty funny... Intuitively it makes sense that machine learning tools would be very trickable since they have no true intelligence. I do think we'll eventually create a true AI with generalized intelligence but not so sure it will happen in our lifetimes.