Face-Recognition

On Tuesday, NBC News reported IBM used photos from Flickr as “training datasets” to help improve its facial recognition technology, unbeknownst to some of the photographers who had taken the images. But lawyers contacted by Legaltech News said photographers and those photographed may have little legal recourse against IBM depending on the specifics of their creative commons license.

Most creative common licenses allow for noncommercial usage, which may raise issues if the images were are used to create a commercial product. 

“Using it internally, you get down to parsing language, and the biggest challenge is what IBM is doing with creating this dataset is something definitely not contemplated by these photographers,” said Phoenix-based business and intellectual property attorney Sara Hawkins. However, “If it's a creative commons noncommercial licence, that presents a problem because IBM is using this for commercial purposes.”

To be sure, copyright issues may arise if the images were reproduced for commercial usage.

“It would be more of a copyright issue in whether there is an appropriate license for IBM to do what it's doing,” said Howard & Howard attorney and University of Detroit Mercy trade law professor Andrew “Jake” Grove. He noted that each copy of the image, through recreating it or reloading the software containing the image, could raise an issue of copyright infringement.

For those photographed, they may also have cause of action based on their right of privacy and publicity. “Those issues are governed by state statutes dealing with right to privacy or state statutes dealing with right of publicity and a federal statute known as Section 43A of the Lanham Act,” Grove said.

However, such cause of actions can be defended by IBM, lawyers said.

“There are two things that IBM could argue. One of them is they have a license because the photographers agreed to let them do it, the other thing is it's copyright fair use,” Grove noted. “It's a defense under the Copyright Act to allow people to use copyrighted works.”

IBM's fair use claim may hinge on the dataset's function as something providing a diverse array of images to improve facial recognition tools. After all, biased coding can have consequential effects in prison sentences to loan acceptance. In turn, industry observers have called for more diverse data for coding algorithms, which IBM claimed was its purpose for the dataset.

“Even if they did have a copyright claim, IBM would have a decent fair use claim, another issue they may rely on,” added Dunlap Bennett & Ludwig partner David Ludwig. “IBM may come out and say, 'We are not copying the photos to sell them, we are using this to create an amazing AI tool.'”

While some photographers may mull over seeking a legal claim, their pursuit may not be fiscally worthwhile. Solo practitioner Carolyn Wright, who counsels photographers, noted most photographers don't register their photos with the U.S. Copyright Office. This allows them to only recoup actual damages, a licence fee usually under $100 and provable profits earned by the company from their image.

It could be difficult to prove what profit a company made from a photographer's image, Wright added. If the image is registered, the photographer could be awarded statutory damages.

While the Flickr photos used by IBM were permitted for public use, collecting public photos on social media would be legally troublesome, especially if the images were used for commercial gain.

“Technically it would be an infringement if it isn't subject to a creative commons license,” Wright said. “Just because I put something on Facebook or Instagram does not mean it can be reproduced.”

As the news of IBM's facial recognition software plays out in the court of public opinion, for Hawkins, the company's use of images represents “uncharted territory” of copyrights of digital images.

“It's really important. We have that data, but the question is how do we allow that data to be accessed and do we allow that data to be accessed for free?” Hawkins said.