Invoke

We are proud to announce our partnership with BRIA AI, bringing their industry-leading visual generative AI models to our platform. Invoke now integrates Bria’s most advanced models which are trained exclusively on 100% licensed data with a full attribution record for every output, fully compliant with the AI Act. Sharing a simple workflow below showing a sketch that is rendered in a specific artistic style using Bria 2.3, their Hard Edge Detection (canny) model, and a custom trained LoRA. Thanks to Andrew Bloom, Polly L., Sharon Dayan and the whole BRIA AI team for making this possible. DM us if you're interested in giving it a go!

Spoke at GDC to an audience with standing room only and a line of 150 out the door, and at the Dell Technologies / NVIDIA DesignAI event at GTC that same night. Generative AI's honeymoon may be over, but there were a lot of people still looking to the horizon to try to grapple with what's coming next. I received some of the most positive feedback I've ever had on this talk specifically, with more than a few artists saying it changed their perspective/stance on AI. My conversations with leaders at major publishers, co-dev/agency partners, and other leaders in the industry all point to on thing. Human-in-the-loop is here to stay. I'm looking forward to sharing the full recording soon! ??

I’ll be taking the stage with Paul Trillo next Monday at SXSW 2025 to talk about AI, creativity, and how to build products that are useful, ethical, secure, and protected. Paul is an incredible director who has been exploring new technologies for creative storytelling since before generative AI was cool. I’m excited for you to see some of the work he has to show. Also - DM me if you’re in Austin and want to meet up! Date: Mar 10, 2025?? Time: 2:30pm – 3:30pm CT? Location: Austin Convention Center,?9ABC #SXSW2025 Link in comments.

I sat down with NBC last week to talk about the work we’re doing to ensure artists can get copyright protection on the works they create in Invoke. Thanks to Gadi Schwartz and his team for spotlighting this historic legal decision (and shout out to Cooley LLP for helping make this a reality). Most companies still think you can’t get a copyright if they use AI at all. Excited to be a small part of helping demonstrate that’s no longer the case!

Join our team!

It's a bit surreal (pun intended), but today Invoke gets to share that we've played a small part in US Copyright history by securing the first copyright for a single image made purely with AI-generated content. Judd Lauter at Cooley LLP states in the article: "What we've demonstrated here is that we've defined a threshold. I think creators are going to continue to test what those boundaries are, both in terms of how human creativity is applied and embodied within content created with the assistance of AI, and also how much creativity is going to be needed to reach that threshold. And the way we figure it out is through example." I've been extremely appreciative of Judd's partnership as we tackled this project, especially because we had to argue diligently with the U.S. Copyright Office (USCO) to reach this point. As we get more clarity for customers on the USCO's stance on AI-assisted works, we'll continue to incorporate leading practices into Invoke. Provenance Records currently captures the entirety of the creative selection/arrangement made by end-users in metadata within Invoke's Canvas, and will evolve later this year to include even more context for the assisted assets produced on platform. This is a big moment for AI-powered creativity, and we're just getting started.

We're excited to have been approached by Daniel Lefcourt at the Rhode Island School of Design for the opportunity to collaborate in teaching these wildly talented students how to use Invoke as an accelerating force in their creative process! Thanks to artnet for telling the story! See comments for a link to the article.

The most controllable AI tooling allows you to change styles, control colors, and transform content easily -- This type of direct control is the future for creatives using these tools. This was a fun workflow that is useful in any serving any type of style transformation need. Some ideas: -- personifying in-game characters for promo materials/key art -- allowing fans to “meet” real-world versions of iconic characters -- exploring how illustrated characters or creative might look in a live-action adaptation -- seeing what a product sketch might look like in real life (eg. on a real person). The key to this workflow is using both a Tile and Canny control and then inpainting specific regions of the original image to regenerate those elements in the new style. DM or comment if you have an image you’d like me to run through this workflow.

Getting AI to nail specific colors in specific areas is hard. With Midjourney, you’re pretty much limited to text prompts, sref codes, generating a bunch of iterations, and crossing your fingers. And even then, you're really lacking a lot of visible control. Sharing a few ways to add colors to specific areas without bleeding into others using Invoke’s canvas: Method #1: Paint color on a raster layer Method #2: Use a tile controlnet. Method #3: Use a raster layer + tile controlnet. Method #4: Use a color palette as a regional reference image. Happy Invoking!

I am honestly more excited than is probably reasonable about the conversation that Bennett and I will be having at SXSW - If our past few conversations are any indication, they'll be raw, philosophical, and full of insight. He's a pioneer, and one of our most creative users - What he's done with Invoke is nothing short of incredible, and I am eager to see the response to his work as he begins to share more of it in 2025.