Home Technology Open-Source Tools The Download: stereotypes of AI models and the new age coding

The Download: stereotypes of AI models and the new age coding

0
The Download: stereotypes of AI models and the new age coding
Meta launches its own AI app
This is the daily edition of The Download

This data set allows researchers to identify harmful stereotypes within LLMs.

So what’s new? AI is rife with culturally specific biases. A new data set called SHADES is designed to help developers tackle the problem by spotting harmful stereotyping and other forms of discrimination that appear in AI chatbot responses in a variety of languages.

What’s the point? While tools to detect stereotypes in AI models exist, they are largely limited to models that have been trained in English. They identify stereotypes within models trained in other language by relying upon machine translations of English, which may fail to recognize stereotypes that are only found in certain non-English dialects. SHADES was developed using 16 languages from different geopolitical regions to avoid these problematic generalizations. Read the complete story.

Rhiannon William

MIT Technology Review Narrated : The second wave in AI coding has arrived

Several startups are racing against each other to create models that produce better and more advanced software. They claim that this is the fastest way to AGI.

We’re publishing a MIT Technology Review narrated podcast every week on Apple Podcastsas well as Spotify. Follow us on either platform and navigate to MIT Technology Review to receive all of our new content.

The must-reads

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.

ChatGPT’s rival 1 Meta has launched a standalone AI app
Meta AI combines an AI assistant and a social media feed. It’s designed primarily around voice conversations. ($Bloomberg )
+It’s sure to be followed by targeted ads. TechCrunch


Jeff Bezos tried to reassure Donald Trump that it was not happening. (WSJ $)
+The market value of Big Tech has plummeted ever since Trump’s inauguration. The Economist ($)
+Tech executives’ loyalty to Trump isn’t being reciprocated. Fast Company ($)

OpenAI has rolled-back an update that made ChatGPT too chatty.
Some users complained it was suddenly becoming too sycophantic. (Ars Technica,)
+Sam Altman admitted the problem. ($Bloomberg )

Huawei is eager to fulfill chip orders from Chinese customers
Since Nvidia has ceased to exist, Huawei is willing to step in. (FT $)
+UK’s semiconductor sector is quietly rebounding. The Conversation


5 The Gates Foundation faces a threat
Due to the Trump administration’s massive cuts in foreign aid, the foundation is struggling. (NYT $)

Six We are living in a new age of deepfake Fraud
The fraudsters are manipulating live video calls. (404 Media)
+A startup created a hyperrealistic deepfake that is so good, it’s frightening. MIT Technology Review

An AI startup made a hyperrealistic deepfake of me that’s so good it’s scary. Citizens in Connecticut are paying for it. (Undark),
+Race to destroy PFAS – the forever chemicals. MIT Technology Review

Eight The number of digital artists in the US has exploded.
This is the fastest-growing segment of the country’s jobs that are internet-dependent. (Axios,)

Fast Company $]

The viral ice-bucket challenge is back
After more than a decade since it first swept across the internet. (WP $)

Quote of the day

“I’m not interested in reading something that nobody said.”

–Emily M Bender, a computational-linguistics professor at the University of Washington, tells the Atlantic why she refuses to use AI text generators.

Another thing

It’s not an easy task. You must first lie motionless in a fMRI scan, possibly for hours, as you watch films or listen audiobooks.

By analyzing the blood flow in your brain, the software can learn to reconstruct what you saw or heard.

Researchers have used generative AI tools like Stable Diffusion or GPT to create more accurate, but not completely accurate, reconstructions of podcasts and films based on neural activity. How close are we to a true “mind reading”? Read the full story.

–Grace Huckins

You can still enjoy nice things

An area for comfort, entertainment and distraction that will brighten your day. (Do you have any ideas? Drop a line (or orskeet them at me).– Finally, some good new–Valerie the miniature Dachshund was rescued after 529 days in Australia’s wilderness. You’re not the only one who has to deal with the constant shouting of “six seven” if you’re a teacher or parent.
How old isMajora’s Mask from The Legend of Zelda ?
+ Meanwhile, in the UK we are turning our front lawns to meadows.

www.aiobserver.co

NO COMMENTS

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Exit mobile version