What is a Deepfake?
A deep fake is A.I. generate media that takes a person’s face and places it upon the body of another person without their consent. Deep fakes became popularized in 2017 and most deep fakes are created as pornagraphy. Other names for this include face swapping. Generally, this is done online using software. There are also specific sites that are generating deep fakes for porn consumption or misinformation. This manipulation of an individual’s face is done with techniques that leverage machine learning and generative artificial intelligence. These models use autoencoders that have been trained by neural network architectures or generative adversarial networks. Deep fakes are most commonly used with celebrities, influencers, and social media personalities. There have also been some instances of political figures being used in deepfakes to promote opposition narratives. This technology is advancing rapidly while legislation and public policy are not keeping up.
Are Deep Fakes Legal?
Currently, there are no Federal laws against creating deep fakes using another individual’s likeness without consent. There was a bill proposed in the 116th Congress House Resolution 3230 (H.R.3230) proposing to require a watermark on media generated using deepfake technology. There is also an act introduced by Congressman Morelle called the “Prevention of Intimate Images Act” which would make it illegal to disclose non-consensual images of an intimate nature that have been altered. Currently, only the following state laws exist regarding the use and creation of deepfakes:
- Texas – A law against the use of deepfakes to influence elections
- Virginia – A law against deep fake pornography
- California – Laws against the manipulative deepfakes within 60 days of an election and non-consensual deepfake pornagraphy
- New York – An amendment to the NDII law that includes “deep fake” images in the definition of “unlawful dissemination or publication of intimate images”
Creators and distributors of deepfakes can potentially be held liable for the creation and dissemination of such material if done so without consent.
How Do You Remove A Deepfake?
We remove deepfakes the same way we remove other harmful information online. We initially contact the creator or distributor of the information for removal, we then contact the platform if needed and lastly we pursue legal means if our initial efforts are not successful.
What Other Types Of A.I. Generated Disinformation Is There?
Our clients have had success removing mugshots from Google for an alleged crime by providing documentation related to the favorable outcome of their case. Court documentation can include any of the following:
- Voice Clones
- Deepfake videos
- Deepfake images
- Synthetic media in the form of audio, video, digital art, photographs etc.
What Can I Do To Protect Myself From Deepfakes On The Internet?
Deep fakes are a violation of one’s privacy, personal data, and human rights. You can become the victim of a deep fake by your photos being used in the generation of an A.I. created deep fake or as a victim of cybercrime or phishing through coercion via deepfaked video of a family member or co-worker.
In order to lesson the chance of your photos being used in a deep fake you should avoid social media or make all of your profiles private. Limit the individuals you allow to access your profiles and avoid placing your face on public forums and dating sites.
If you ever receive a video of a family member, co-worker, or friend requesting money, login credentials, account numbers, PINs or other sensitive information double check with the person and verify that it is indeed them making the request.