Watermarking AI Content: Microsoft and OpenAI Agree

by time news

The big companies of AI they support the California law to put watermark on contents

Microsoft, OpenAI, Adobe and also the owner of X Elon Musk would support the law of the American state of California AB 3211, which requires the presence of watermarks in content produced by artificial intelligence. Several companies had shown hostility to this law in the spring, when it was initially drafted. After some amendments, however, it seems that big tech has changed its mind.

AB 3211 It is the first law that attempts to directly regulate the products of generative artificial intelligence so as to make them clearly distinguishable from real ones. The means by which it would like to achieve this result would be the watermark, the transparent symbols that are often used to determine the ownership of a certain image and prevent its piracy on the Internet.

In this case, however, the law does not only provide graphic symbols. The most important change in the text is the one regarding metadata. Each image and each video carries with it additional information to that which makes up the images. This is data that allows us to identify, for example, the place or date in which a photo was taken or the device with which it was made.

AI-generated images and videos currently have no standard for this data. AB 3211 iwould require companies to create one, so that anyone accessing the metadata could immediately notice the artificial origin of the product.

One problem that the law does acknowledge, however, is that few people know about or can access metadata when viewing an image on the web. For this reason, a second part of the law would require social networks to clearly indicate content generated with artificial intelligence. Elon Musk has decided to support this legislative initiative, a sign that even from this point of view the text could obtain developments.

……… KLEINROCK REGISTRATION

You may also like

Leave a Comment