Saturday, January 28

DALL-E works with the images of creators who do not receive anything in return: what copyright says about AI


The images of DALL-E are amazing, but the AI ​​is not starting from scratch. If they can create them, it is because they have been trained with millions of images from the internet. What this AI does is base itself on the images that match the key elements of the description that we give it.


Sometimes the resulting image is extremely similar to the original photo. This has sparked debate again: isn’t the AI ​​infringing the copyrights of the original cartoonists or creators of the photo? The answer is not simple, because the World Intellectual Property Organization (WIPO) itself is reflecting on how to adapt copyright laws to the rise of artificial intelligence.

That picture looks a lot like mine. Some artists they have already started complaining. Tools like DALL-E Mini take advantage of network images to generate others, but they do so in an opaque way and without the possibility of establishing to what extent the original image has been used.

The evidence is often obvious, but for an artist it is practically impossible to determine how the process has been. That is why they have shouted to the heavens and request that AIs comply with copyright, despite the fact that at the moment it does not seem that companies like OpenAI are having much concern in this field.

Also Read  These resistance exercises will help you sleep, and they are better than cardio

An AI cannot be considered an author. And therefore there is no infringement. As explained by WIPO, copyright laws establish that only one person can be the author of a creative work. In Spain, this is expressly included in the Intellectual Property Law: “A natural person who creates a literary, artistic or scientific work is considered an author.” Therefore, the creations of a machine or an AI such as OpenAI are not considered an infringement, because there is no authorship.

The same position was adopted in 2019 by the US Copyright Office, establishing that “works produced by a machine or a mere mechanical process that works randomly or automatically without any input or creative intervention from a human author will not be registered”.

Taking advantage of this legal situation, the big technology companies are training their machines with large databases without worrying about possible intellectual property conflicts. The matter is far from being closed. Nerea Sanjuán, a lawyer in the Commercial Law Area of ​​Uría Menéndez, explains that “the possible legal solutions that arise in this matter include revolutionary solutions that would confer the status of “author” of the work to the machine itself”.

Is there creativity or is it just a good combination? The Google engineer who spoke to LaMDA was convinced that the AI ​​had a conscience. It’s hard to believe. However, it is a much more earthly debate whether AIs like DALL-E are creative and original. That is to say: the resulting images are a creative work or simply a complex combination of other images. It’s a subtle difference, but it’s what ultimately determines the copyright debate.

To determine if it is a derivative work we should “see how the AI ​​works, how it reformulates or transforms the work, how it learns and creates new elements and is based on patterns that it sees work. It would have to be seen case by case and the exceptions applicable”, explained to Xataka, Sergio Carrasco, jurist of Fase Consulting.

Also Read  This was the Mercury, the amazing retro-futuristic train that circulated through the streets of New York for 23 years

The problem is that by not being able to audit the code of these AIs, establishing how it works is much more complicated. Unfortunately, this issue is not even being discussed in the future Artificial Intelligence legislation of the European Union.

GitHub Copilot and the controversy with the copyright: the debate on whether AI infringes the copyright of code written by other programmers

DALL-E puts the scope of the Creative Commons license on the ropes. The training of tools like DALL-E Mini is based on open datasets, one of them being YFCC100M, which contains some 99.2 million photos from Flickr, all of them shared under various Creative Commons licenses.

But the massive and indiscriminate use of these images does not fit well with the CC license. An equivalent case is the one that occurred with Github Copilot, where Julia Reda, the only representative of the Pirate Party in the European Parliament, explained that “those who argue that it is a derivative work of the training data hope that this production will be published under the same GPL license “. I mean, if they use a Creative Commons image of mine, I hope the AI ​​image will also be released under this license.

Microsoft offers something that OpenAI does not yet: a duplication detector. The difference between some AIs and others is a matter of complexity, because the operation in the end is the same. Github Copilot is an AI that learns from other people’s code, while OpenAI’s DALL-E is an AI that learns from other people’s images. In the case of Microsoft, to tackle this dilemma, they began to offer a duplication detector. A tool that offered the possibility to “include proper attribution or decide not to use the code at all.”

Also Read  In search of the best quality-price sound bar (2022): purchase recommendations and seven models from 140 euros

However, this option might not be feasible in OpenAI due to a matter of complexity and difficulty in determining to what extent an image has been relevant.

The solution, if there is one, goes a long way. We are at the gates of knowing what the implications of these tools based on artificial intelligence are. Despite the fact that this issue is being debated within the World Intellectual Property Organization, there is no concrete initiative for the application of a change.

A possible solution that some experts propose is the creation of a newly created copyright for massive use by AI, for specific cases. An addition that should be part of future legislation and would greatly affect copyright management.

It would also be necessary to see if the European Union and the United States see it in the same way. It doesn’t seem like we’re going to have this debate any time soon, as work is still being done today to properly regulate technologies that came along years ago, such as social networks.

In Xataka | I have asked the DALL-E Mini to show Steve Jobs eating an apple and I have just understood why this AI is essential



Leave a Reply

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