Photobucket, once the world’s leading image-hosting site, has undergone a significant decline in its user base over the years. Currently, only 2 million people continue to utilize the platform, according to Similarweb. However, there may be a glimmer of hope for Photobucket in the form of the generative AI revolution.
CEO Ted Leonard, who oversees the operations of the company with a team of 40 from Edwards, Colorado, has disclosed that he is engaged in discussions with multiple tech companies. The purpose of these negotiations is to license Photobucket’s vast collection of 13 billion photos and videos for training generative AI models. These models have the capability to produce new content based on text prompts.
Regarding the pricing structure, Leonard has mentioned that rates range from 5 cents to US$1 per photo and more than US$1 per video. The variations in prices depend on the buyer’s requirements and the specific types of imagery desired. In fact, Leonard shared that some prospective buyers have expressed a need for over a billion videos, a number surpassing the available content on the Photobucket platform.
Photobucket has refrained from disclosing the identities of potential buyers due to commercial confidentiality. However, these ongoing negotiations, which have not been previously reported, suggest that the company possesses content that could potentially be worth billions of dollars. This sheds light on the emergence of a thriving data market driven by the race to dominate generative AI technology.
Tech giants like Google, Meta, and Microsoft-backed OpenAI initially relied on freely scraped internet data to train generative AI models, such as ChatGPT, which are capable of emulating human creativity. Although these companies argue that this practice is both legal and ethical, they have faced legal action from copyright holders for copyright infringement.
Simultaneously, these tech companies have discreetly been paying for access to content that is locked behind paywalls and login screens. This has given rise to a hidden trade in various types of content, including chat logs and forgotten personal photos from obsolete social media applications.
Edward Klaris from law firm Klaris Law, which advises content owners on licensing deals for AI training, emphasizes the pursuit of copyright holders with private collections that are not publicly available for scraping. According to Klaris, deals worth tens of millions of dollars have been made to license archives of photos, movies, and books.
In an effort to delve deeper into this nascent market, Reuters conducted interviews with over 30 individuals familiar with AI data deals, including industry executives, legal experts, and consultants. The insights gained from these discussions shed light on the types of content being acquired, the prices involved, and emerging concerns pertaining to personal data being used in AI models without individuals’ knowledge or explicit consent.
Although major market research firms have yet to estimate the size of the opaque AI data market, some, including Business Research Insights, estimate it to be approximately $2.5 billion presently, with forecasts suggesting it could reach nearly US$30 billion ($45.6 billion) within the next ten years.
The pursuit of data for generative AI models coincides with the mounting pressure faced by developers of these models to account for the massive quantities of content used for training. This process, known as “training,” necessitates substantial computing power and can take several months to complete. Tech companies argue that the use of freely scraped web page data is crucial for making the technology economically viable.
While Photobucket’s user base may have significantly declined, the potential value of its vast collection of photos and videos in the generative AI market highlights new possibilities for the company. As the generative AI revolution unfolds, the acquisition and utilization of data become key factors in driving innovation and advancements within the field.
FAQs
1. What is generative AI?
Generative AI refers to the technology that enables machines to generate new content, such as images, videos, or text, based on input prompts or training from existing data.
2. What is the current user base of Photobucket?
According to Similarweb, Photobucket currently has around 2 million users.
3. How much does Photobucket charge for licensing its photos and videos?
Photobucket offers pricing ranging from 5 cents to US$1 per photo and more than US$1 per video, depending on the buyer’s requirements and the type of imagery sought.
4. Are there concerns about personal data being used in AI models without consent?
Yes, there are emerging concerns about personal data being utilized in AI models without individuals’ knowledge or explicit consent. This raises ethical and privacy considerations in the development and use of AI technologies.
5. How much is the estimated size of the AI data market?
While the size of the opaque AI data market has yet to be accurately determined, some estimates, such as those provided by Business Research Insights, suggest it is currently valued at approximately $2.5 billion, with projections indicating potential growth to nearly US$30 billion ($45.6 billion) within the next decade.
Photobucket, once a leading image-hosting site, has experienced a decline in its user base over the years, with only around 2 million users currently utilizing the platform. However, the company may find hope in the generative AI revolution.
CEO Ted Leonard has disclosed that Photobucket is in discussions with multiple tech companies to license its vast collection of 13 billion photos and videos for training generative AI models. These models have the ability to create new content based on text prompts. The pricing structure for licensing Photobucket’s content ranges from 5 cents to US$1 per photo and over US$1 per video, depending on buyer requirements and the type of imagery desired. Some potential buyers have expressed interest in over a billion videos, surpassing the available content on the Photobucket platform.
While Photobucket has not revealed the identities of potential buyers, these negotiations suggest that the company possesses valuable content potentially worth billions of dollars. This highlights the emergence of a thriving data market driven by the race to dominate generative AI technology.
Major tech giants like Google, Meta, and Microsoft-backed OpenAI initially relied on freely scraped internet data to train generative AI models. However, these companies have faced legal action from copyright holders for copyright infringement, despite arguing that the practice is legal and ethical. As a result, they have been discreetly paying for access to locked content behind paywalls and login screens, creating a hidden trade in various types of content.
Content owners with private collections that are not available for scraping have been pursued for licensing deals. Deals worth tens of millions of dollars have been made to license archives of photos, movies, and books. This nascent market raises concerns about personal data being used in AI models without individuals’ knowledge or explicit consent.
While market research firms have not yet estimated the size of the opaque AI data market, Business Research Insights approximates it to be around $2.5 billion presently. However, forecasts suggest that it could reach nearly US$30 billion ($45.6 billion) within the next decade.
The acquisition and utilization of data are becoming key factors in driving innovation and advancements within the generative AI field. Despite Photobucket’s reduced user base, its extensive collection of photos and videos holds potential value in the generative AI market, presenting new opportunities for the company.
The source of the article is from the blog aovotice.cz