The Meta Controversy

Today:

Protesters Decry Meta’s “Irreversible Proliferation” of AI

Meta's been catching some heat lately for openly sharing its AI tech, with some folks comparing it to spreading dangerous weapons. Last week, a bunch of protesters camped outside of Meta's San Francisco digs, saying that this open approach to AI is kinda like letting the cat out of the bag - once it's out, there's no putting it back in.

Holly Elmore, who's been sounding the alarm on AI safety, led the charge. She points out that while companies like OpenAI and Google keep their AI stuff more under wraps, Meta just put theirs out there for everyone. This could be bad news because, in the wrong hands, these tools can be twisted to do some sketchy stuff.

READ THE ARTICLE ON IEEE SPECTRUM

Altana using AI to map and analyze global supply chain with launch of next-gen Atlas

AI startup Altana has released a new tool, Atlas, that provides a dynamic map of the global supply chain. Based in New York, Altana partnered with U.S. Customs, Maersk, and other big names to tackle issues like forced labor and ensure trade law compliance using AI. Their newest Atlas version taps both public and private data to chart the supply chain, featuring an AI assistant that understands and answers questions in everyday language.

The idea? As goods move worldwide, they generate tons of data. Altana's Atlas uses AI to make sense of it all, showing how products move from raw material to finished good. It's like Google Maps, but for products. This helps companies see and handle potential supply chain hiccups, like unexpected events or natural disasters.

READ THE ARTICLE ON FOX

The big AI and robotics concept that has attracted both Walmart and Softbank

Softbank made some bucks from its deal with Arm Holdings, but now it's eyeing a new AI prize: Symbotic, a Massachusetts-based software and robotics firm. This isn't news to Walmart, as they've already grabbed a piece of that pie.

Symbotic's robots are designed for warehouses. Their tech aims to make warehouses smarter, faster, and more efficient. With Softbank, they're launching GreenBox Systems, aiming to give smaller businesses a taste of this high-tech warehouse action. Experts think this could be a game-changer, making fancy warehouse tech affordable, kinda like how cloud computing made high-end tech available for everyone.

READ THE ARTICLE ON CNBC

Artists across industries are strategizing together around AI concerns

Artists are raising the alarm about AI taking over creative jobs. Whether it's fake songs by popular artists or AI-drawn profile pics, tech's influence in art is growing fast. To tackle this, a rights group and a music labor organization started a campaign, #AIdayofaction. They want Congress to stop big companies from copyrighting AI-generated art, to keep humans in the loop.

There's talk that artists from different fields are not communicating enough about their shared worries. The goal? Make sure AI helps individuals earn more without being taken advantage of by big corporations. It's a tricky topic; artists fear being replaced, while music platforms use AI more and more.

Universal Music wasn't happy when an AI-made song imitating their top artists went viral. At a recent discussion, the FTC brought together creative folks to chat about AI's impact. They agree: while tech can help make art, it can't replace the human touch in creativity.

READ THE ARTICLE ON TECHCRUNCH

AI firms working on “constitutions” to keep AI from spewing toxic content

Big AI companies are beefing up their chatbots, but they're also trying to make sure these bots don't go rogue. Microsoft's OpenAI can now chat using voice, pictures, and text, while Facebook's Meta is rolling out AI celebs for folks on WhatsApp and Instagram. But, there's a catch. As these bots get smarter, it's getting trickier to keep them from spouting nonsense or even harmful stuff.

To tackle this, big shots like Google DeepMind and Anthropic are penning "AI constitutions" – basically rulebooks that teach the bots to behave. It's like giving them a code of ethics. Right now, a lot of companies have humans review the bot's answers to teach them what's cool and what's not. Another method is to have experts try to trick the system to expose flaws.

READ THE ARTICLE ON ARSTECHNICA

AI can detect if you have recently smoked cannabis

Scientists whipped up an AI that can tell if you've just toked up, aiming to quickly figure out if someone's too high and might need medical help. Traditional tests, like pee and spit tests, take ages because pot doesn't show up right away in some body juices.

So, Sang Won Bae's crew at a New Jersey tech school had 33 regular pot users wear a fancy wristband and tracked their phone wiggles to see how steady they were. Using this data, they trained the AI, and it nailed if someone had just blazed 85% of the time.

READ THE ARTICLE ON NEW SCIENTIST

Fact check: AI doctors on social media spreading fake claims

Some "doctors" on social media aren't the real deal. Videos show them in lab coats, spouting health advice. But they're just computer-generated bots, not real MDs, and some of what they're saying is straight-up false.

A lot of these fake videos are coming out of India, where there's been a big spike in health misinformation, especially during the COVID-19 times.

These AI "doctors" look real convincing. They wear the right clothes, and they sound super smart. The goal? To make you trust them, maybe even buy something.

READ THE ARTICLE ON DW

AI: A Catalyst in Astronomical Progress

AI is like a supercharged telescope for modern astronomers. It's becoming a must-have tool, making sense of space data and helping to spot cool new stuff in the universe.

For example, AI helps to clear up space pictures, making them sharper. It even played a big part in getting that epic shot of a massive black hole a while back. Also, when it comes to big mysteries like dark matter and dark energy, AI is on the case. Using fancy computer simulations, it helps scientists get a grip on these hard-to-understand parts of the cosmos.

And for sorting and naming space stuff? AI's got that too. It's making sure that when scientists categorize stars, planets, and other objects, they're doing it right and fast.

READ THE ARTICLE ON NEAT PROMPTS

Pause For Thought: The AI Pause Debate

Last month, a bunch of big brains sat down to chat about whether we should hit the brakes on AI development. Despite everyone agreeing AI could be a potential global disaster, they couldn't agree on what "pausing" meant, if it could work, or even if it was a good idea. The gist of their debates was:

• Simple Pause: Should AI companies chill for a bit, like 6 months or so? Most folks don't think this will do much. Wait too long, and technology could jump forward too fast and get unpredictable. Plus, if some countries pause and others don’t, the ones chillin’ might fall behind.

• Surgical Pause: This idea tweaks the simple pause. The questions are: When should we pause? And for how long? The trick is figuring out the right timing so the "good guys" keep their lead but also get a chance to make AI safer.

• Regulatory Pause: What if we just waited until we set up a watchdog agency (like the FDA for AI)? That way, we could monitor the big, new AIs for danger signs. Some folks argue that these types of agencies usually slow down progress a lot, but hey, at least it'd be safe, right?

• Total Stop: Some believe we should halt AI entirely. But that's a tough sell. How do you enforce it globally? And as tech keeps getting better, it’s like trying to keep water in a sieve.

• No Pause: Others argue, “Let's just keep rollin'!” Pausing might just create more problems like illegal AI labs or giving other countries a head start. Plus, it could be tough to decide when to unpause.

Despite being called a "debate", not many folks went toe-to-toe. But the comment section on one post got spicy with discussions about whether treaties could work and how history might inform our AI choices.

READ THE ARTICLE ON ASTRAL CODEX TEN