Reports indicate that Photobucket, a popular platform for hosting images and videos, is currently in discussions with artificial intelligence companies specializing in generative technology to obtain licenses for using its vast library of content for training algorithms.
Once a prominent player in the internet arena with 70 million users during the early 2000s, Photobucket now counts merely two million users. Nevertheless, the emergence of generative AI has presented a potential opportunity for the Colorado-headquartered company.
According to CEO Ted Leonard in an interview with Reuters, discussions are underway with numerous tech firms to license the platform’s impressive collection of 13 billion photos and videos. These assets are intended for training text-to-image algorithms, made possible by the company’s revision of its terms of service last October to grant itself the authority to commercialize any uploaded content for AI training purposes.
Leonard emphasizes the necessity to generate revenue to sustain operational costs and potentially provide continued support for free user accounts through these licensing agreements.
He elaborates on negotiations surrounding pricing models, citing potential rates ranging from 5 cents to $1 per photo and exceeding $1 per video, with variations to accommodate the preferences of different buyers and the specific imagery they require. Details regarding the interested parties remain undisclosed by Leonard.
Reflecting on these discussions, Leonard reveals encounters with buyers expressing the need for substantial amounts of content, surpassing Photobucket’s current library size. He expresses astonishment at the scale of demands, questioning the feasibility of such requests.
The revelation that Photobucket explores the option of selling its users’ content to generative AI companies coincides with industry reports highlighting data scarcity challenges faced by tech giants like OpenAI and Google in sourcing materials for AI applications.
Comparisons drawn by Reuters liken the current situation to a frenzied “gold rush,” fueled by legal considerations regarding the accessibility of data, particularly publicly available data from sources like the non-profit entity Common Crawl. Major players in the tech sector, including Meta, Google, Amazon, and Apple, have forged partnerships with Shutterstock to leverage its extensive database, with deals rumored to range from $25 million to $50 million.
However, concerns exist within the industry regarding the procurement of content from platforms like Photobucket, characterized by murky licensing rights. Daniela Braga, CEO of Defined.ai, an entity that licenses data to prominent tech corporations, voices apprehensions, highlighting the risks associated with utilizing unverified content in AI applications. Braga emphasizes potential ethical dilemmas that may arise if AI generates unauthorized representations of individuals without proper consent.
