Googles latest AI video generator can render cute animals in implausible situations

AI-generated video is still in a primitive state, but it has been progressing in quality over the past two years. In October 2022, we covered Google’s first publicly unveiled image synthesis model, Imagen Video. It could generate short 1280×768 video clips from a written prompt at 24 frames per second, but the results weren’t always coherent. Before that, Meta debuted its AI video generator, Make-A-Video. In June of last year, Runway’s Gen2 video synthesis model enabled the creation of two-second video clips from text prompts, fueling the creation of surrealistic parody commercials.

• Tom Glitter and Sparky Buttons owe their cute names to a neural network.
• And Hugging Face is set to release its open-source version of GPT-3 in the next couple of months.
• A place for comfort, fun and distraction in these weird times.
• “A video about a kangaroo trying to check in to an aeroplane—it fundamentally doesn’t really matter whether you engage with the idea that’s a deepfake or not.
• Essential digital access to quality FT journalism on any device.

FT recommends

The exact place AI will have in society in the future is a conversation that will continue for years to come. The line preventing that reality shouldn’t have to move just because it’s now suddenly cute. Social media has several niches, and cute animals are one of them. Whether it be Pudgy Penguins, Milk & Mocha Bear, or Tonton Friends, their creators spend days and weeks sketching and writing scenes for their accounts. “A video about a kangaroo trying to check in to an aeroplane—it fundamentally doesn’t really matter whether you engage with the idea that’s a deepfake or not.

“The stock image industry is officially toast,” tweeted another. Google Brain has revealed its own image-making AI, called Imagen. AI content creators threaten these artists by posting similar content that is just as or even more popular, and for not even a fraction of the effort and talent put into making organically-made art. Officials ultimately decided not to reimburse her, but her professor did end up expressing regrets about the situation.

He’s currently working on his second novel, a science-fiction book. Mr Elon Musk’s sassy “anti-woke” chatbot Grok is a tech-person term for understanding something (as in, you grok a design). —An open letter from 42 Democrat lawmakers urges Google CEO Sundar Pichai to stop collecting location data that could be used to identify people seeking abortions, according to CNBC. We’re starting to give AI agents real autonomy, and we’re not prepared for what could happen next. It’s now possible to run useful models from the safety and comfort of your own computer. OpenAI is making DALL-E 2 available only to a handful of trusted users; Google has no plans to release Imagen.

Grok, o3 and ELMo — there’s a reason AI names are so weird

Shane has become the go-to person for entertaining neural net names. She previously trained AI systems to come up with cookie names (Hallowy Maples, Apricot Dream Moles) and D&D monsters (Spectral Slug, Jabberwont). She originally trained a network by giving it 8,000 cat names to learn from, but the AI stumbled, giving out names like Hurler and Retchion. Shane revisited the challenge with an AI that had a lot more context and trained it up with a long list of cat names. Despite the nerves and preparation, Mills said he loves getting the opportunity to read graduates’ names. He even came back last year during his sabbatical for the role.

Several other universities, from the University of South Carolina to the University of Georgia, said they would employ similar technology this spring. “Slop” is the term for low-quality, AI-generated images and videos, and they can often go surprisingly viral. Some of the people sharing the clips may not even realize the videos are AI-generated, though we hope that goes without saying when you come across a video of baby Ice Cube (Lil’ Ice Cube?) rapping “It Was a Good Day.” Another month, another flood of weird, wonderful and cute images generated by an artificial intelligence. In April, OpenAI showed off its new picture-making neural network, DALL-E 2, which could produce remarkable high-res images of almost anything it was asked to. Mixing concepts like “fuzzy panda” and “making dough” forces the neural network to learn how to manipulate those concepts in a way that makes sense.

Students are asked to mark up the card with guidance on how to pronounce their names and turn them in the Friday before commencement. Since 2019, Mills, chair of Otterbein University’s Department of Philosophy and Religion, has held the unenviable role of reading hundreds of graduates’ names before they walk across the commencement stage each spring. To some at the university, he’s affectionately called “The Voice of God.” A presenter at Thomas Jefferson University made headlines last spring after she accidentally mispronounced the names of nursing graduates during a commencement ceremony.

The following year, that former student found him before the ceremony started and said her younger sister was graduating that day, so she wasn’t going to have him repeat his mistake from the previous year. Though it might seem self-explanatory to practice running through his list of names, Mills said an experience he had early on as the commencement reader solidified his routine. Incidents like what happened at Thomas Jefferson last year are worst-case scenario, but many students at institutions experimenting with AI solutions say they are willing to risk mistakes to have a human touch. Pace isn’t alone in swapping live talent for machine precision, but for many students, the experience felt more “Black Mirror” than commencement celebration.

The professors are using ChatGPT, and some students aren’t happy about it

Designed to repel cheap Chinese rivals, it chalked up another victory for the sector’s mystifying inability to think up coherent names. But these firms are pushing the boundaries of what AI can do and their work shapes the kind of AI that all of us live with. They are creating new marvels, but also new horrors—and moving on with a shrug. When Google’s in-house ethics team raised problems with the large language models, in 2020 it sparked a fight that ended with two of its leading researchers being fired.

Read the full story—and see more pictures created by Imagen—here. Bova, who has also worked as a correspondent for The Boston Globe, wrote about his disdain for the AI tool in a recent op-ed for The Huntington News, Northeastern’s independent student newspaper. By opting to have AI read the names instead of a real person, the university “showed more reverence for artificial intelligence than for its student body,” Bova wrote. Google’s brand new Veo 3 video generator is currently the most advanced video generator, but it also costs $250 a month to access as part of Google’s AI Ultra Subscription. I was able to create a fairly realistic baby video using Veo in Gemini, but that’s not exactly the point of the meme.

Trending Now

A complete version of the Wave will go on sale in the fall, which means that users itching for a duck-flavored Champ should get access to the full extent of Clova’s features the moment they go on sale. Here’s hoping Google responds with a version of the Google Home that looks like a bugdroid — Mountain View’s smart speaker will touch down in Japan by the end of the summer. Lumiere can also do plenty of party tricks, which are laid out quite well with examples on Google’s demo page.