Tech

Artificial intelligence discovers how to make up its own creepy language, confusing researchers

AN ARTIFICIAL INTELLIGENCE program has developed its own language and nobody can understand it.

OpenAI is a developer of artificial intelligence systems – their programs are fantastic examples of super computing, but there are quirks.

An AI system assigned these gibberish names to objects, and researchers don't know why

1

An AI system assigned these gibberish names to objects, and researchers don’t know why

DALLE-E2 is OpenAI’s latest AI system – it can generate realistic or artistic images from user-entered text descriptions.

DALLE-E2 represents a milestone in machine learning – OpenAI’s website says the program “learned the relationship between images and the text used to describe them”.

A DALLE-E2 demonstration contains interactive keywords that visitors can play with and generate images. Switching between different keywords results in different images, styles, and themes.

But the system behaves strangely – it writes its own language out of random arrangements of letters, and researchers don’t know why.

The AI ​​face search engine triggers fears when it finds pictures of you on the internet
Could artificial intelligence REALLY wipe out humanity?

Giannis Daras, a computer science graduate student at the University of Texas, published a Twitter thread Details on the unexplained new language of DALLE-E2.

Daras asked DALLE-E2 to create an image of “farmers talking about vegetables” and the program did so, but the farmers’ speech read: “vicootes” – an unknown AI word.

daras fed “vicootes” back into the DALLE-E2 system and got back pictures of vegetables.

“We then feed the words: ‘Apoploe vesrreaitars’ and we get birds.” Daras wrote on Twitter. “It seems the farmers are talking about birds messing around with their vegetables!”

Daras and a co-author have written a paper on DALLE-E2’s “hidden vocabulary”.

They concede that asking DALLE-E2 to generate pictures of words – the command “a picture of the word plane” is Daras’ example – usually results in DALLE-E2 spitting out “gibberish text”.

When reconnected to DALLE-E2, this gibberish text leads to images of airplanes – which says something about how DALLE-E2 is talking to and thinking about itself.

Mysterious death of star collegiate lacrosse player who died'unexpectedly' at age 19
I'm a Sports Illustrated Swimsuit Model - my way of preparing for my shoot

Some AI researchers argued that DALLE-E2’s gibberish text was “random noise.”

Hopefully we won’t find out that the DALLE-E2’s second language was a security hole that had to be patched after it’s too late.

We pay for your stories!

Do you have a story for The US Sun team?

https://www.the-sun.com/tech/5478378/artificial-intelligence-spotted-inventing-its-own-language/ Artificial intelligence discovers how to make up its own creepy language, confusing researchers

Chris Barrese

InternetCloning is an automatic aggregator of the all world’s media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials, please contact us by email – admin@internetcloning.com. The content will be deleted within 24 hours.

Related Articles

Back to top button