Positive, the rating of a soccer recreation is necessary. However sporting occasions can even foster cultural moments that slip below the radar—comparable to Travis Kelce signing a coronary heart to Taylor Swift within the stands. Whereas such footage might be social-media gold, it’s simply missed by conventional content material tagging programs. That’s the place Twelve Labs is available in.
“Each sports activities workforce or sports activities league has a long time of footage that they’ve captured in-game, across the stadium, about gamers,” Soyoung Lee, co-founder and head of GTM at Twelve Labs, informed Observer. Nonetheless, these archives are sometimes underutilized as a consequence of inconsistent and outdated content material administration. “Up to now, a lot of the processes for tagging content material have been guide.”
Twelve Labs, a San Francisco-based startup specializing in video-understanding A.I., needs to unlock the worth of video content material by providing fashions that may search huge archives, generate textual content summaries and create short-form clips from long-form footage. Its work extends far past sports activities, touching industries from leisure and promoting to safety.
“Giant language fashions can learn and write rather well,” stated Lee. “However we wish to transfer on to create a world through which A.I. can even see.”
Is Twelve Labs associated to Eleven Labs?
Based in 2021, Twelve Labs isn’t to be confused with ElevenLabs, an A.I. startup that focuses on audio. “We began a 12 months earlier,” Lee joked, including that Twelve Labs—which named itself after the preliminary dimension of its founding workforce—typically companions with ElevenLabs for hackathons, together with one dubbed “23Labs.”
The startup’s bold imaginative and prescient has drawn curiosity from deep-pocketed backers. It has raised greater than $100 million from buyers comparable to Nvidia, Intel, and Firstman Studio, the studio of Squid Sport creator Hwang Dong-hyuk. Its advisory bench is equally star-studded, that includes Fei-Fei Li, Jeffrey Katzenberg and Alexandr Wang.
Twelve Labs counts hundreds of builders and tons of of enterprise prospects. Demand is highest in leisure and media, spanning Hollywood studios, sports activities leagues, social media influencers and promoting companies that depend on Twelve Labs instruments to automate clip era, help with scene choice or allow contextual advert placements.
Authorities businesses additionally use the startup’s expertise for video search and occasion retrieval. Past its work with the U.S. and different nations, Lee stated that Twelve Labs has a deployment in South Korea’s Sejong Metropolis to assist CCTV operators monitor hundreds of digicam feeds and find particular incidents. To cut back safety dangers, the corporate has eliminated capabilities for facial and biometric recognition, she added.
Will video-native A.I. come for human jobs?
Most of the industries Twelve Labs serves are already debating whether or not A.I. threatens people jobs—a priority Lee argues is simply partly warranted. “I don’t know if jobs will likely be misplaced, per se, however jobs should transition,” she stated, evaluating the shift to how instruments like Photoshop reshaped artistic roles.
If something, Lee believes programs like Twelve Labs’ will democratize artistic work historically restricted to corporations with massive budgets. “You at the moment are in a position to do issues with much less, which implies you have got extra tales that may be created from unbiased creatives who do not need that very same capital,” she stated. “It really permits for the scaling of content material creation and personalizing distribution.”
Twelve Labs shouldn’t be the one A.I. participant eyeing video, however the firm insists it serves a unique want than its a lot bigger opponents. “We’re excited that video is now beginning to get extra consideration, however the way in which we’re seeing it’s numerous innovation in giant language fashions, numerous innovation in video era fashions and picture era fashions like Sora—however not in video understanding,” stated Lee, referencing OpenAI’s text-to-video A.I. mannequin and app.
For now, Twelve Labs provides video search, video evaluation and video-to-text capabilities. The corporate plans to broaden into agentic platforms that may not solely perceive video but additionally construct narratives from it. Such fashions might be helpful past artistic fields, Lee stated, pointing to examples like retailers figuring out peak foot-traffic hours or safety purchasers mapping the sequence of occasions surrounding an accident.
Whereas A.I. would possibly assist a Hollywood director assemble a film, Lee believes it gained’t ever be the director. Even when the expertise can present narrative choices, people nonetheless determine which story is most compelling, determine gaps and provide the footage. “On the finish of the day, I feel there’s nothing that may exchange human artistic intent.”
