Meta introduces generative AI video advertising tools

Last Week in AI: Google, OpenAI leads generative AI video creation

Getting the prompt to produce an AI output aligned with what the creator envisioned is a core goal, and eschewing overt aesthetics for Veo is critical to achieving a wide-ranging adaptability. Some 30 percent of the developers who responded to the survey said they felt negatively about AI, opposed to 18 percent last year; only 13 percent believed AI was having a positive impact on games, down from 21 percent in 2024. “No matter how you put it, generative AI isn’t a great replacement for real people and quality is going to be damaged,” another developer wrote in their response.

With Firefly, you can now generate videos using either a text-based prompt or by uploading an image for it to work with. Whichever option you pick, Adobe will limit you to five seconds of video, so it’s best for creating short clips instead of entire videos. It may be that we need an entirely new approach; or it may be that Gaussian Splatting (GSplat), which was developed in the early 1990s and has recently taken off in the image synthesis space, represents a potential alternative to diffusion-based video generation.

By blending innovation with accessibility, Meta seems poised to redefine the future of video production. Users can generate stunning HD videos (up to 1080p resolution) lasting up to 16 seconds, all by typing a few instructions. The AI model, built with 30 billion parameters, accounts for intricate details like camera movement, object interaction, and even physical laws, resulting in content that feels natural and immersive. Veo creates videos with realistic motion and high quality output, up to 4K. Explore different styles and find your own with extensive camera controls.

New GeForce RTX 50 Series GPUs Double Creative Performance in 3D, Video and Generative AI

The practical use of various API-based generative video systems reveals similar limitations in depicting accurate physics. However, certain common physical phenomena, like explosions, appear to be better represented in their training datasets. YouTube will also integrate generative AI text and image output into an “Inspiration” feature for creators, which is intended to feed them suggestions and examples for video content. Generative AI will even offer “AI-enhanced” suggestions for how creators can respond to comments. An agent that can get data from the internet in real time can book your flights and hotel rooms, make dinner reservations, and more, based on what it knows about you and your upcoming travel—all without your having to guide it.

Mark Zuckerberg Unveils Game-Changing AI Video Creation in an Unexpected Demonstration – Jason Deegan

Mark Zuckerberg Unveils Game-Changing AI Video Creation in an Unexpected Demonstration.

Posted: Sat, 25 Jan 2025 13:00:57 GMT [source]

Successful deployment of generative AI requires trust across your institution’s journey, from partners to data, to models to outcomes. Hooglee appears to be the first artificial intelligence project that Schmidt has personally incubated after investing in a number of AI companies, such as Anthropic and quantum computing startup SandboxAQ. The billionaire, who Forbes estimates is worth more than $26 billion, has also funded an OpenAI grantmaking program and the AI science nonprofit FutureHouse. Instead, what they do is automate many of the tasks involved in video editing, such as cutting footage to insert, re-ordering scenes, adding audio effects, generating captions, and adding avatars, graphics, or animations.

Adobe makes generative AI video model safe for enterprises

In the first example of a group of friends sitting on the trunk of a car, the original prompt includes mention of “flash photography,” but the subjects are clearly backlit. One could argue that a flash was used to create intense backlighting, but if the idea behind the prompt was to create something representative of flash photography from the 1960s, this image isn’t it. With this research release, we have made the code for Stable Video Diffusion available on our GitHub repository & the weights required to run the model locally can be found on our Hugging Face page. Further details regarding the technical capabilities of the model can be found in our research paper. We’re continuing to improve our generative AI models and look forward to helping build technologies that unlock creativity for everyone, everywhere.

• A group of researchers, including interns at Netflix, have developed of Go-with-the-Flow, a new model that aims to provide an easy way to control motion patterns in video diffusion models.
• Drag-centric applications have become frequent in the literature lately, as the research sector struggles to provide instrumentalities for generative system that are not based on the fairly crude results obtained by text prompts.
• Some 30 percent of the developers who responded to the survey said they felt negatively about AI, opposed to 18 percent last year; only 13 percent believed AI was having a positive impact on games, down from 21 percent in 2024.
• Amid a mix of cultural and economic factors impacting the industry, developers are also still dealing with company enthusiasm for technology that some find ethically concerning.

Descript is something of a disruptive player in the generative video market as a startup launched by one of the co-founders of Groupon. By gaining a reputation as a powerful tool for social media content creators and TikTok kids, it’s hit a marketing sweet spot, but it’s also used across a wide range of corporate and educational use cases. Synthesia is a generative AI video tool that is particularly great when it comes to creating digital avatars and bringing them to life with realistic voice and animation. It’s used by businesses to create training materials as well as by marketers for promotional and advertising content, and it would also be a great choice for educational videos. The majority of surveyed developers (52%) said they work for companies that use generative AI, with 36% reporting that they personally use such tools for game development. While the majority of surveyed developers are working at companies that use generative AI, 30% of respondents said they believe generative AI is having a negative impact on the video game business (up 12% from 2023).

According to Meta, the Movie Gen collection is made up of four models that enable video generation, personalized video generation, precise video editing and audio generation. Ali said the AI-generated Veo creations will be watermarked with SynthID, a tool developed by DeepMind for watermarking and identifying AI-generated material, and will be applied with a label that clearly communicates to viewers that it was generated with AI. As AI-generated content has flooded YouTube, Google and the rest of the internet, most of it has not featured watermarks, a trust and safety commitment made by many companies that researchers have found is easy to bypass. Today ChatGPT is able to generate responses for very current news events, as well as near-real-time information on things like stock prices.

Language models have a tendency to make stuff up—they can hallucinate nonsense. Moreover, generative AI can serve up an entirely new answer to the same question every time, or provide different answers to different people on the basis of what it knows about them. OpenAI’s ChatGPT now has access to the web, making it far better at finding up-to-date answers to your queries. The startup Perplexity was doing the same, but with a “move fast, break things” ethos. Literal trillions of dollars are at stake in the outcome as these players jockey to become the next go-to source for information retrieval—the next Google.

Inspired by the NotebookLM podcast feature, I set out to build an application that would convert my articles to video talks. The key step is to prompt an LLM to produce slide contents from the article, another GenAI model to convert the audio script into audio files, and use existing Python APIs to put them together into a video. I didn’t do this, but you can squint a little and see where things are headed. If I’d wrapped up the presentation creation, audio creation, and movie creation code in services, I could have had a prompt create the function call to invoke these services as well. And put a request-handling agent that would allow you to use text to change the look-and-feel of the slides or the voice of the person reading the video.

WIPRO Executive Chairman Rishad Premji, who is at Davos, told NDTV that the generative artificial intelligence (AI) is the big focus now. AI tools are getting better at these things, so it’s worth expanding your horizons. Check the patterns on objects like art and clothing — it may look bad, arbitrary, or repetitive under closer scrutiny.

The surveyed developers cited things like IP theft, energy consumption, and AI program biases as contributing to their feelings toward generative AI. Today, we’re bringing our new Veo 2 capabilities to our Google Labs video generation tool, VideoFX, and expanding the number of users who can access it. We also plan to expand Veo 2 to YouTube Shorts and other products next year.

And so in 1994 Jerry Yang created Yahoo, a hierarchical directory of websites. TBH, and with the benefit of hindsight, I think we all thought it was much better back then than it actually was. This website is using a security service to protect itself from online attacks. There are several actions that could trigger this block including submitting a certain word or phrase, a SQL command or malformed data. GeForce RTX 5090 and 5080 GPUs will be available for purchase starting Jan. 30 — followed by GeForce RTX 5070 Ti and 5070 GPUs in February and RTX 50 Series laptops in March. The NVIDIA Blueprint for 3D-guided generative AI is expected to be available through GitHub using a one-click installer in February.

Women and non-binary developers account for 32% of all game developers, up from 29% and 24% in the prior years. The percent of respondents identifying as LGBTQ+ was 24%, and white/caucasian developers accounted for 59% of the total game developer base in the study. Since prior approaches to the problem do not offer drag-based editing, the researchers opted to compare Framer’s autopilot mode to the standard functionality of older offerings. ‘Specifically, we concatenate the VAE-encoded latent feature of the first [frame] with the noisy latent of the first frame, as did in SVD. Additionally, we concatenate the latent feature of the last frame, zn, with the noisy latent of the end frame, considering that the conditions and the corresponding noisy latents are spatially aligned. The new system, titled Framer, can not only follow the user-guided drag, but also has a more conventional ‘autopilot’ mode.

Meta rolled out generative AI tools to resize images, change backgrounds and repurpose existing images to create multiple versions of ads in May. Since then, brands have been able to use these capabilities when building out their campaigns on Meta Ads Manager. MovieGen is Meta’s latest offering in the realm of generative AI, designed to take visual storytelling to a whole new level. According to Meta’s Chief Product Officer, Chris Cox, while the technology isn’t yet available to the public, early results have shown that MovieGen outperforms existing competitors such as Runway Gen-3, Luma Dream Machine, and OpenAI Sora. The system excels in creating fluid, lifelike movements, a significant step forward in AI-generated visuals.

Whether you’re crafting a quirky social media video or working on a Hollywood-level production, this AI tool promises to make video creation more accessible and flexible than ever. At its core, MovieGen is a multimodal AI model that can create high-quality videos and audio content based on user prompts. It allows for detailed customization, whether it’s adding effects, altering costumes, or incorporating props—all controlled via simple text commands. Former Google CEO Eric Schmidt has spent the last few months working on an artificial intelligence project that aims to capitalize on the booming landscape of AI video generation.

In a blog post, Meta’s AI team explained that it’s aiming to usher in a new era of AI-generated content for creators on its platforms. The Meta Movie Gen models build on the company’s earlier work in generative AI content creation, which began with its “Make-A-Scene” models that debuted in 2022, enabling users to create simple images and audio tracks, and later videos and 3D animations. Meta’s later Llama Image foundation models expanded on this work, introducing higher-quality images and videos, as well as editing capabilities. MediaTek and Kuaishou have collaborated to bring generative video (Gen-AI) capabilities to smartphones powered by Dimensity 9300 and Dimensity 8300 chipsets. Kuaishou’s image-to-video technology leverages the powerful 7th generation NPU in the Dimensity mobile platform to produce dynamic videos based on image prompts directly on the smartphone.

“Hey Google, make me a show about a six-year-old that won’t clean up her room, who gets lost under a pile of undies and has to earn the trust of the mice and spiders to form a coalition and plan an escape.” Gaál has used Veo to pump out footage that might have cost millions and the work of a decent-sized team to shoot in the real world – illustrating that it’s probably not a great time to be getting into the film industry. And the short clips used in this piece are a stylistic choice rather than a reflection on the quality of the AI’s longer scenes. “It’s more about feeding the story [than trying to hide crappy AI generation],” writes Gaál. One of the tasks it specializes in is automatically creating highlight reels showcasing the most important, exciting or interesting moments from longer videos. It then makes it simple to quickly repurpose them as trailers or social clips to engage more viewers.

3D artists have adopted generative AI to boost productivity in generating draft 3D meshes, HDRi maps or even animations to prototype a scene. At CES, Stability AI announced SPAR3D, its new 3D model that can generate 3D meshes from images in seconds with RTX acceleration. Streamlabs Intelligent Streaming Assistant is an AI agent that can act as a sidekick, producer and technical support.