2nd Set AI Launches to Help Entertainment, Media and Sports Enterprises Create and Deliver Generative Images and Video for Global Audiences - While Protecting IP and Elevating Brands

Backed by Cortical Ventures, Firsthand, Company Ventures and angel investors, the Company also shared proprietary research validating widespread addressable markets

NEW YORK, Aug. 13, 2025 /PRNewswire/ -- 2nd Set AI, a company providing AI solutions for enterprises looking to deploy generative images and video to enhance creativity while protecting IP and brand guidelines, today announced its formal launch along with pre-seed funding led by Cortical Ventures with participation from Firsthand.vc and Company Ventures as well as a collection of AI expert angels. The new capital will allow 2nd Set AI to accelerate hiring, build new business relationships and open up an Early Access Program to help top entertainment organizations generate images and videos from their valuable IP.

The company also released proprietary research around accuracy, responsibility, and efficiency of image generation. Its team studied the major image generation models in use in industry and found substantial issues with inaccuracy, particularly in more complex compositions. Aside from broad inaccuracies and off brand content, the research found a range of issues with unauthorized replication (IP piracy), toxicity, and bias. Finally, it found that the above issues combine to create a much higher fully loaded cost of deploying generative images once the above issues are taken into account.

Founded in 2024 by AI veterans Jeffrey Smith and Saurav Pandit, 2nd Set AI offers a suite of generative visual capabilities for enterprises, chiefly IP rights holders and licensees. Its platform allows entertainment and sports media companies to use images of people and characters for which they control the rights to generate new images and videos from those owned images. Uniquely, the 2nd Set AI platform is constantly checking and revising its own work to ensure the highest quality outputs available, greatly exceeding the accuracy of solutions provided by even much larger companies. This enables traditional media companies to use generative media with confidence that generative media produced from their valuable IP will honor their business goals and protect their brand equity.

"When founding 2nd Set, we recognized that this was a huge moment for the impact of generative media on traditional media sectors like entertainment, advertising and marketing, stock media, and more," said Smith, CEO & co-founder.  "But as practitioners, we recognized that there was a huge mismatch between the outcomes that were promised and the extant models on the market. Most providers were greatly overpromising their capabilities and users were getting frustrated and disillusioned. One major advertising organization we talked to was producing 96 image contact sheets to cherry pick one good image, and that was after a human reviewed all of them! We recognized that there was a major missing component in the generative media stack: a multimodal reasoning model that understood business requirements. We serve organizations seeking to transform existing imagery they already own or license and use it in new ways using generative visual models."

In his prior role, Smith spent six years at Facebook/Meta where he managed the PyTorch team, and led key acquisitions and research. He remains an active angel investor and LP in venture capital funds. Pandit has worked for years at the intersection of AI and creativity, including NLP and Computer Vision AI, and in the world of stock media at Pond5 and Shutterstock. Smith and Pandit have partnered before in two previous startups, Intent Media (still in operation as Black Crow Labs) and John Done, an early vocal agent startup.

"One thing that stands out from our research is that there is not one single model that is good at doing everything, and is well behaved in every possible way," said Pandit, chief science officer & co-founder.  "For example, for superior spatial layout and scene coherence, Open AI's GPT-4o, with its powerful native LLM reasoning, seems to be a good choice. But it collapses with an increasing number of object attributions in the prompt. Conversely, Google's recent Imagen models are more robust to increasing the count of things, but don't get spatial relationships. The flagship 4 Ultra version of Imagen performs better than Imagen 4 in some ways, but of course it costs more. And both of them struggle generating crowds.  The newest Flux Kontext models provide balanced performance, but are heavily guard-railed and aggressively block prompts.So you can see that choosing the right set of AI tools for your task could require your org to have a team of AI engineers constantly onboarding and testing new image and video gen models - which by the way come out every other week! Our team knows this pain because we lived it!"

2nd Set AI is already partnering primarily with several large IP rights holders, primarily across entertainment, sports, and licensed media, across large bodies of content including images, logos, videos, comic character libraries, even entire sports leagues. The company has rolled out an intelligent agent that allows IP rights holders to request a report – delivered via an in-depth analysis and shareable infographic – that illustrates how often their IP is generated by various commercially available generative models. IP rights holders who would like to check for unauthorized replication of their owned imagery are encouraged to try it by going to this page or email it directly at [email protected].

"Having come up through the technical ranks in the early stages of the AI revolution and with years of startup and big tech experience at companies including Meta and Shutterstock, Jeff and Saurav are exactly the kind of founders we look to partner with," said Jeremy Achin, cofounder and managing partner, Cortical Ventures. "This is a massive challenge not solvable without the right technical underpinnings. In contrast to overly optimistic and risky generative AI technologies from other companies, 2nd Set has developed a platform that reasons throughout the entire generative process to ensure accuracy and compliance."

About 2nd Set AI

Founded by AI experts from Meta and Shutterstock, 2nd Set AI is a dedicated team of researchers and creative technologists in New York City, backed by experienced AI investors including Firsthand.vc, Cortical Ventures, and Company Ventures, as well as notable AI angel investors. For more information, visit 2ndset.ai.

About Cortical Ventures

Cortical Ventures is a venture capital firm focused on helping entrepreneurs building the next generation AI Companies. The firm was started by DataRobot founder Jeremy Achin and Igor Taber, who previously led Corporate Development at DataRobot and was an early investor in the company while he was at Intel Capital. Cortical Ventures was started to invent, incubate and invest in the companies leading the AI revolution. The firm is backed by leading VC firms and partners, AI luminaries and top founders and operators in the industry.

Press Contact:
Tim Turpin
[email protected] 

SOURCE 2nd Set AI