- Deep fake will be a threat to courtroom evidence submissions.
- Russia, China and North Korea could use it to embellish their military capabilities.
- The neural rendering can be used to force a fake image to give scripted speeches using anyone's voice.
Now everyone in the western world has access to high speed internet and a 4G network. On the global scale, a majority of nations have access to it as well. The evolving technology allows quality videos and images in high definition (HD). Social media is one of the biggest news sources for many and one of the most powerful social engineering tools in use.
The biggest threat from this technology is “deep fake” (named after deep learning and fake) that models the behavior of any living being on video. The base for the technology is the Generative Adversarial Network (GAN), a class of machine learning systems invented by Ian Goodfellow and his colleagues in 2014. Two neural networks contest with each other in a game. Given a training set, this technique learns to generate new data with the same statistics as the training set.
The algorithms in the program constantly compete with each other by learning from the images to create real faces and copy them many times until the machine gets confused over which is the real image vs the fake one. Recently deep fake technology was applied to generate a fake video of former US President Barack Obama using profanities against current US President Donald Trump.
Originally, deep fake has been mostly used to create celebrity impersonations, but these days it can be used to generate propaganda for social engineering purposes. It is plausible to believe that during the next US election deep fake technology will be one of the issues we face. Russia can be identified as one of the countries interested in deploying the technology, perhaps by the previously exposed troll farm operating out of St. Petersburg. Additionally, Ukraine has been in the news a lot recently, where there are interests and groups with deep ties to certain high profile individuals who are set up to conduct election meddling.
Buy Research Report: Global Fresh Milk Market Research Report 2021-2025 – Groupe Lactalis, ADM, CHS, Manildra Group
Buy Research Report: Global Natural Mineral Water Market Research Report 2021-2025 – Danone, Bongrain, Devondale Murray Goulburn, Fonterra
Get Content Links from Established Websites
I Give You 23 Million Verified Emails
Buy Research Report: Serial Console Server Market to Be Valued at US$ 37 Bn by 2030 – Supply Issues amidst COVID-19 Pandemic Impeding Market Expansion
The next level of technology is Neural Rendering. It was designed with collaboration by Princeton, Stanford and the Max Plank Institute. It allows you to force an image to recite any text using any person’s voice. All it needs is a 40 minute video of the person’s recorded voice. The technology will copy the mannerisms and the voice, transforming them into an image of the person speaking the desired text.
One of the biggest challenges we’ll face with these developing fake technologies is in the courtroom. From a legal standpoint, it will create an issue for both sides. The prosecutors would face significant issues about whether or not they can rely on video surveillance. However, the defense might also attempt to submit a fake video.
Russia, China and North Korea could use this technology to embellish their military might with orchestrated videos and machines which are nonexistent or have much less power than projected, distributed through their propaganda channels.
It is certain such technology is going to be one of the biggest threats to us in many arenas and will forever change the norms of evidence and public skepticism.