Adobe could analyze its users’ creative processes to train its AI, artists say

The web community of artists is increasingly afraid of artistic production tools based on artificial intelligence. According to some, Adobe could track creatives with its software to analyze behavior and thus feed its own artificial intelligence.

Tools for creating images from text made possible thanks to artificial intelligence have been met with great success for several months on the Internet and above all, they are increasing: DALL-E 2, Midjourney, etc. . Tech giants are also taking action, such as Meta, which manages to create videos from text. For its part, OpenAI, known as ChatGPT, is developing AI capable of building 3D models.

To work, this software needs a lot of data, beyond the complexity of development or the computing power, it is the courage of war for developers. Currently, automatic image generation tools are based on… the Internet, which is full of artistic works. But this is not necessarily done with the permission of the creators and some automatically created images have been denounced for plagiarism on social networks. The media Fast Company has reported that some artists are accusing Adobe of stalking their creative process to train its artificial intelligence. An accusation that Adobe denies.

An option enabled by default in Creative Cloud that allows Adobe to use its users’ data

A few days ago, comic book author Claire Wendling posted a screenshot of an Adobe privacy and personal data setting on Instagram. It indicates that “Adobe may analyze your content using techniques such as machine learning […] to develop and improve its products and services“. A setting enabled by default that allows Adobe to review User Content.for product improvement purposes“. Such a parameter exists in many other software and online services.

The fear shared by many artists, especially on Twitter, is that software from the Adobe suite ” track artists who use their apps to see how they work, meaning they steal processes and actions that graphic designers have developed over decades of work to exploit in their own automated systems“, summaryFast Company. In other words, they fear that their complex artistic processes can be studied by an artificial intelligence, which can automate them. Something to replace their work one day.

Adobe says it doesn’t use user data to develop AI

An Adobe spokesperson said the company does not use customer information:Adobe does not use any data stored in customers’ Creative Cloud accounts to train its experimental generative AI features“. In addition to this, Adobe said it is reviewing its policy “to better define use cases for generative AI“. However, the service’s FAQ states that “Adobe may analyze your Creative Cloud or Document Cloud content to provide product functionality and to improve and develop our products and services“. This includes all files stored on Adobe’s servers.

Results for “An astronaut riding a horse in a photorealistic style” // Source: OpenAI

However, we will find out later that “None of your content is included in our products or services unless you make it public […] The information obtained through content analysis is not used to reproduce your content or to identify any personal information.Moreover, according to Michael Velea, a professor at University College London who specializes in digital rights,Under European privacy law, Adobe must obtain user consent before reading data from individuals’ devices for purposes not necessary for the service requested by the user.“. He added that some creative professionals may sign nondisclosure agreements with clients: Adobe’s setting, which is enabled by default, could cause legal issues there.

A sense of dread that is bigger than the Adobe question

As mentionedFast CompanyThe “The fear that design processes will be monitored by the technology and then used to train artificial intelligence, is part of a wider anxiety about how artists are treated by generative AI applications.“.

An explanatory diagram about Dall-E // Source: OpenAI

Midjourney, for example, did not seek permission from the artists it had “suckedworks to do so, which were refuted by the artist community. Some are, moreover,create tools for their colleagues to opt out of using their works to train AI“. This use would violate the intellectual property and privacy rights of millions of creators. Moreover, the European Union questions the legality of automatic generation tools of AI, from the point of view of the protection of data.

To follow us, we invite you to download our Android and iOS application. You can read our articles, files, and watch our latest videos on YouTube.

Leave a Reply

Your email address will not be published. Required fields are marked *