OpenAI recently introduced GPT-4o, its latest AI model that incorporates textual, vocal, and visual functionalities. Touted as a major improvement over its predecessors in speed and functionality, GPT-4o is offered to users for free. However, this free service comes with a hidden cost: users’ data.
Also Read: OpenAI Unveils ChatGPT-4o: Free Access with Limitations
“It’ll be free for all users, and paid users will continue to have up to five times the capacity limits of free users.”
OpenAI CTO Mira Murati
AI tool seizes third-party content, raises privacy concerns
GPT-4o collects extensive user data, including texts, audio recordings, and images. This data collection goes beyond the user’s data to the data of other parties involved in the interaction. For instance, when a user uploads a screenshot of an article from the New York Times for summarization, OpenAI holds the copyrighted content within the screenshot. This practice is also quite problematic from the point of view of privacy and copyright, as users are often unaware that they are sharing data that does not belong to them.
Also Read: Has Safety Taken a Back Seat at OpenAI?
“GPT-4o gobbles up not only users’ own information but also third-party data that are revealed during interactions with the AI service.“
The Strategist
OpenAI embraces “surveillance capitalism,” profits from data
Due to the scarcity of suitable training data, the quality of training data is also decreasing. Consequently, OpenAI has used several approaches to acquire data. As the New York Times reported, OpenAI had transcribed more than a million hours of YouTube videos, which was against YouTube’s rules in late 2021.
While Google, the owner of YouTube, did not file a lawsuit, possibly because it also collects data, this case shows how far AI companies are willing to go to get data.
Source: Hype Insights
In particular, with GPT-4o, OpenAI quickly accumulates a large amount of user-generated multimodal data, following a similar approach to we-media business models. This model is where technology platforms offer products and services for free and earn money through collecting user data. This approach reflects the general trends of “surveillance capitalism,” in which the firms benefit from the surveillance of the targets.
Although users have consented to such activities, copyright infringement questions remain relevant, The Times stated. This can lead to the sharing of data that users do not possess, thus creating problems for content creators. Outputs from AI are often unrecognizable from the original dataset, making it hard for the copyright owner to spot an infringement.
Content creators take action, ban AI’s use of their work
The use of AI for unauthorized training is a real problem, and content creators and publishers are starting to take protective measures. Some use technological measures to prevent data scraping, while others change the terms of service to ban the use of their content in making AI. Recently, Sony Music issued letters to more than 700 generative AI firms and streaming services with a cease and desist order.
Also Read: Prominent Musicians Demand Safeguards Against AI’s Encroachment
Moreover, more than 200 musicians, including Billie Eilish, Katy Perry, and Smokey Robinson, have signed an open letter to AI developers, tech companies, and digital platforms. In the letter, they demand that the companies “cease the use of artificial intelligence (AI) to infringe upon and devalue the rights of human artists.”
As previously reported by Cryptopolitan, ChatGPT was down early on Tuesday, and users were unable to access the innovative artificial intelligence tool for several hours. The outage began at 00:PDT and was mainly heard from the mobile user community, especially those who use the ChatGPT Android application. According to DownDetector, most issues arose within the ChatGPT product rather than the website or the mobile application interface
Cryptopolitan reporting by Brenda Kanana