DALL·E 3 Now on ChatGPT Plus & Enterprise

Today:

DALL·E 3 is now available in ChatGPT Plus and Enterprise

DALL·E 3 is now in ChatGPT Plus and Enterprise. They've put safety measures in place to stop it from making sketchy images, like violent or inappropriate ones. They tested it with some users who gave 'em tips on where the blind spots were. They've also tried to make sure it doesn't copy living artists or show famous faces, and they want the pictures to represent everyone fairly. If users see something off, they can give feedback using a flag icon. The team's also cooking up a tool that can spot if a picture was made by DALL·E 3 with a pretty high accuracy, even if the picture's been tweaked a bit. But it's not foolproof. They're hoping it'll help folks figure out if they're looking at AI-made stuff in the future. It's a work in progress!

READ THE ARTICLE ON OPENAI

YouTube might make an official way to create AI Drake fakes

YouTube's working on a new tool to mimic famous voices like Drake's using AI. They've chatted with music companies to get permission to use their tunes for this AI tool, but no big names have said "yes" yet. Last month, YouTube rolled out some other AI features, but couldn't add this voice clone thing because they didn't have the rights.

This whole voice cloning thing became a hot topic when a fake "Drake" song blew up online. Some artists are cool with it, but others like John Legend and Selena Gomez aren't too thrilled and want rules. It's kinda murky right now whether this AI voice thing breaks any copyright laws.

Other big tech companies have also jumped on the AI music bandwagon. YouTube says they want to help the music world figure out this new tech, and the music companies seem on board. But the big question: Can YouTube legally let folks use these voice tools without getting slapped with lawsuits?

READ THE ARTICLE ON THE VERGE

NVIDIA Expands Robotics Platform to Meet the Rise of Generative AI

NVIDIA's beefing up its robotics game big time to handle the growing demands of this thing called Generative AI, which is like giving superpowers to AI models to understand stuff better and faster. Basically, it's the new hotness in tech that's helping industries from robot pals to delivery drones do their jobs more efficiently.

They're rolling out some fancy updates to their NVIDIA Jetson platform, which is what developers use to make AI and robots smarter. Two main parts of this platform, the NVIDIA Isaac ROS (for robots) and NVIDIA Metropolis (for edge AI), are getting these big upgrades.

What's super cool is this new Jetson Generative AI Lab they're launching. It's a spot for developers to play around and experiment with the latest AI models. For the tech-savvy peeps, there's a bunch of tools and updates to help make AI applications even better. And for the robot nerds, NVIDIA's also giving robots better "eyes and ears" with improved software, so they can move and react better.

READ THE ARTICLE ON NVIDIA

Microsoft's new AI assistant can go to meetings for you

Microsoft is launching a new AI buddy called Microsoft 365 Copilot on November 1st. This assistant can sit in on your Teams meetings and give you the highlights so you don't have to show up. It can also whip up emails, PowerPoints, and other documents real quick.

There are rules saying you gotta know if you're chatting with a machine or a person. Microsoft says it's on the user to spill the beans about using AI, but the European Union thinks companies should ensure that's happening.

The tech that powers this assistant is from OpenAI, which Microsoft has poured a lot of cash into. During a sneak peek, the assistant was seen swiftly handling emails and making PowerPoint slides. But, if you ask it for some flair in a presentation, it's a no-go. It's pretty sharp in breaking down Teams meetings though. However, it won't dish on who rocked or bombed in a meeting.

READ THE ARTICLE ON BBC

DeepMind Wants to Use AI to Solve the Climate Crisis

DeepMind, that Google-owned AI powerhouse, is hustling to use its tech brainpower to tackle climate change. They've got a three-pronged game plan:

• Understand the Mess: AI can make sense of the climate chaos by predicting and monitoring changes better.

• Fix What We Got: Instead of ditching all the old systems, AI can soup 'em up to be more efficient.

• Invent Cool Stuff: AI's not just about killer robots and space-age tech from the movies. It can push the envelope on new discoveries. DeepMind's nerding out over nuclear fusion, a fancy way to get loads of clean energy, and they're using AI to figure it out.

But there's a catch. AI needs power – a lot of it. So folks are wondering if AI's carbon footprint is worth the benefits. DeepMind's take? They're working on making AI solutions more carbon-efficient.

READ THE ARTICLE ON WIRED

Amazon unveils new AI-powered technology that will inspect delivery vans and help keep drivers safe

Amazon's got a new trick up its sleeve: a fancy AI tech that checks out delivery vans for any little issues before they turn into big problems on the road. This gizmo, called Automated Vehicle Inspection (AVI), is like a doctor's checkup but for vans. After a day's work, drivers roll their vans through this "scanner" which quickly spots and classifies problems. The tech even spots things like tire issues – a major cause of on-road headaches.

The cool part? The more it's used, the smarter it gets. And it's not just for vans – in the future, it might be checking out bikes, drones, and even the inside of vehicles. For fleet owners like Bennett Hart, this means less time poking around vehicles and more time making sure deliveries run smoothly.

READ THE ARTICLE ON AMAZON

‘Mind-blowing’ IBM chip speeds up AI

IBM in San Jose has cooked up a new computer chip that's kinda like a brain. It's super fast for AI tasks, doesn't waste a lot of power, and doesn't need to keep grabbing memory from other places. This means it can recognize stuff in pictures way faster and with less juice than other chips.

People are saying it's a game changer. This NorthPole chip uses something called neural networks, which are like layers of mini-computers working together to spot patterns. Usually, chips have to keep reaching out for more memory, which slows everything down. This slowdown's been a big headache for computers for ages and uses a ton of energy.

But this NorthPole chip? It's got 256 mini-computers, and each one has its own memory. They're connected in a way that's inspired by our brain, and it just crushes other AI machines in tests. It's also way less thirsty for power. If they made it with the latest tech, it'd be even better.

READ THE ARTICLE ON NATURE

NVIDIA AI Now Available in Oracle Cloud Marketplace

NVIDIA's AI tech is now up on Oracle's online store. So, if you're using Oracle's cloud, you can now get NVIDIA's super-duper AI computer power with just a few clicks. This is a game-changer because it's the first time something like this has been done. Loads of companies use Oracle's cloud for their apps – we're talking big names in healthcare, research, finance, and more. With NVIDIA in the mix, they can now train AI for stuff like chatbots and content creation.

A university in New York (University at Albany) is already using this new NVIDIA tech in Oracle's cloud to teach and research AI topics, from health to computer chip design. They're all about making sure their students are ready for future jobs. NVIDIA's offering also comes with tools for cybersecurity, computer vision, and other AI-related tasks.

READ THE ARTICLE ON NVIDIA

AI-powered lasers could zap space debris away from collision courses

Space junk's a real headache, especially with all the satellites and important stuff like the International Space Station up there. To tackle this, some smart folks are thinking about using AI-controlled lasers in space to gently push any threatening junk into safer paths. NASA's giving this idea a thumbs up with a $200,000 grant over three years.

The idea is to have multiple lasers working together because one laser just won't cut it. Even though we're tracking a lot of big junk pieces, there are millions of smaller ones we can't keep an eye on. These lasers could be a game-changer, helping to prevent costly delays, like when Amazon and India had to push back their satellite and moon launches because of space junk.

READ THE ARTICLE ON SPACE

AI is resurrecting the dead, with or without permission

Some folks are using fancy tech tools to whip up videos of the dearly departed, even if they never got the okay. This could be a comfort to some, but it's also kicking up dust about privacy and saying "yes" before you shuffle off. For example, Robin Williams' daughter Zelda ain't happy seeing AI-made videos of her dad. She finds them creepy.

This isn't the first rodeo for bringing celebs back, though. We've had movie biopics and even hologram concerts of artists who've left the building, like Elvis. But experts are raising their eyebrows, wondering what kind of can of worms this AI thing might open. Like, who's going to keep these tech systems going, and what does it mean for those trying to move past their grief?

On the flip side, some folks dig it. Kim K was all heart-eyes over an AI hologram of her late dad. And over in China, they're using AI to help folks cope with loss.

READ THE ARTICLE ON AXIOS

Watch Mark Zuckerberg use AI Ray-Bans to guide him through how to braid his daughter's hair

Mark Zuckerberg just gave hair YouTubers a run for their money by using Meta's latest Ray-Ban smart glasses to braid his daughter's hair. Sharing an Instagram video, Zuckerberg is seen chatting up the glasses' AI, asking how to do a braid. Once he's done, he snaps a pic to show his wife, Priscilla. Though he proudly captioned the post "Finally learned to braid. Thanks, Meta AI", the tech mogul had a bit of a tangle with the hair tie at the end, admitting braiding ain't really his strong suit.

These snazzy glasses, launched at the 2023 Meta Connect, come with video and audio recording and are powered by Meta AI. By just speaking up, you can ask the AI questions or share your view on Facebook and Instagram.

READ THE ARTICLE ON BUSINESS INSIDER

Evaluating social and ethical risks from generative AI

You know how AI is now writing books, making designs, and even helping doctors? Cool stuff, right? But hold up. We gotta make sure these AI robots don’t go rogue on us. We gotta keep them in check!

So, some folks wrote this paper, and they’re like, "Let's break down checking AI into three parts." First, we see what the AI can do. Second, we see how people use it. And third, we see what happens when a lot of people use it in the real world.

They found out that right now, most tests only look at what AI can do, but that ain't enough! We also need to see how AI might mess with our heads or the bigger picture, like our jobs and society. Like, an AI might be great at spitting out facts, but what if it gets those facts wrong, and people start believing lies?

Here's the deal: building AI that's safe isn't just on the guys making it. Everyone's gotta pitch in. The AI makers, the app builders, the government, and even us regular Joes and Janes. It's like a big team huddle.

READ THE ARTICLE ON DEEPMIND