That is at the moment’s version of The Download, our weekday e-newsletter that gives a each day dose of what is going on on on this planet of expertise.
Catching unhealthy content material within the age of AI
Large Tech is surprisingly unhealthy at catching, labeling, and eradicating dangerous content material. In principle, new advances in AI ought to enhance our capability to do this. In follow, AI isn’t excellent at deciphering nuance and context. And most automated content material moderation techniques have been skilled with English information, which means they don’t perform properly with different languages.
The latest emergence of generative AI and enormous language fashions like ChatGPT signifies that content material moderation is more likely to turn out to be even more durable.
Whether or not generative AI finally ends up being extra dangerous or useful to the net info sphere largely hinges on one factor: AI-generated content material detection and labeling. Read the full story.
Tate’s story is from The Technocrat, her weekly e-newsletter providing you with the within observe on all issues energy in Silicon Valley. Sign up to obtain it in your inbox each Friday.
Should you’re fascinated about generative AI, why not try:
+ spot AI-generated textual content. The web is more and more awash with textual content written by AI software program. We want new instruments to detect it. Read the full story.
+ The within story of how ChatGPT was constructed from the individuals who made it. Read our exclusive conversations with the important thing gamers behind the AI cultural phenomenon.
+ Google is throwing generative AI at the whole lot. However specialists say that releasing these fashions into the wild earlier than fixing their flaws may show extraordinarily dangerous for the corporate. Read the full story.
Podcast: Harvesting the long run with AI and satellites
Within the newest episode of our award-winning podcast, In Machines We Belief, the group travels from check farms to labs to learn the way AI is utilized in farming in some methods you won’t anticipate, like for monitoring the well being of crops—from house. Listen on Apple Podcasts or wherever you get your podcasts, and take a look at the first installment of this mini sequence in case you haven’t already.
I’ve combed the web to search out you at the moment’s most enjoyable/necessary/scary/fascinating tales about expertise.
1 Chatbots are extra influential than you understand
They’re in a position to subtly alter our opinions in an identical approach to different people. (WSJ $)+ Immediately’s chatbots are simply the tip of the AI iceberg. (Wired $)+ ChatGPT goes to vary training, not destroy it. (MIT Technology Review)
2 Russia’s plane are stuffed with US expertise
Despite the fact that sanctions are supposed to stop US items from getting into the nation. (NYT $)
3 Gory information footage is just ever a couple of clicks away on social media
However it’s up for debate whether or not seeing it serves any actual function. (WP $)
4 Sam Altman’s Worldcoin is nearing $100 million of funding
Regardless of some fairly main privateness and moral points. (FT $)
+ Study concerning the dodgy strategies Worldcoin used to recruit its first half one million check customers. (MIT Technology Review)
5 College students are the US semiconductor trade’s secret weapon
Corporations want extremely expert engineers, and universities are pleased to oblige. (IEEE Spectrum)
6 Why it’s so tough to wash up aviation’s carbon footprint
Demand for flights is just rising. (Economist $)
+ How new applied sciences may clear up air journey. (MIT Technology Review)
7 What’s taking place with Google Assistant, precisely?
The once-ubiquitous AI helper was noticeably absent from its I/O convention. (Wired $)
+ Google is prompting us how you can immediate its personal AI fashions. (The Verge)
+ That wasn’t Google I/O — it was Google AI. (MIT Technology Review)
8 Fireplace-suppressing chemical substances are poisonous to wildlife
A milestone lawsuit may cease emergency providers from utilizing it sooner or later. (The Atlantic $)
+ The hunt to construct wildfire-resistant houses. (MIT Technology Review)
9 Meet the web’s favourite metallic detectorists
The fun of unearthing buried treasure is rife amongst on-line communities. (The Guardian)
10 Amazon is attempting to construct a wiser house robotic 🤖
If it really works, it’ll be a souped-up model of its Astro robotic with ChatGPT-style smarts. (Insider $)
Quote of the day
“Should you can survive right here you may make it wherever.”
—Tech entrepreneur Rosie Zhang tells the Financial Times how China’s intensely aggressive surroundings helps its native corporations to strike it huge abroad.
The massive story
That is how AI bias actually occurs—and why it’s so exhausting to repair
If we wish to have the ability to repair bias in AI, we have to perceive the mechanics of the way it arises within the first place.
We frequently shorthand our clarification of AI bias by blaming it on biased coaching information, however the actuality is extra nuanced. Bias can creep in lengthy earlier than the information is collected in addition to at many different levels of the deep-learning course of—and might be extremely exhausting to repair. Read the full story.
We are able to nonetheless have good issues
A spot for consolation, enjoyable and distraction in these bizarre occasions. (Obtained any concepts? Drop me a line or tweet ’em at me.)
+ Should you’re having fun with The Legend of Zelda: Tears of the Kingdom, right here’s a list of magical movies that flatter its vibe completely.
+ Be nonetheless my beating coronary heart: Dolly Parton is moving into rock.
+ California’s superbloom is admittedly very stunning.
+ These adorable baby goats aren’t simply cute—they’re additionally achieved firefighters (thanks Anna!)
+ A bit reminder of the significance of the ‘third’ places we spend time in exterior of labor and residential.