|Artificial intelligence is leading to advancement after advancement. Aumcore, a globally integrated multidisciplinary marketing agency based in New York, reported that the global AI market is projected to increase by $89.9 billion in 2025 / Photo by: tiagozr via 123RF|
Artificial intelligence is leading to advancement after advancement. Aumcore, a globally integrated multidisciplinary marketing agency based in New York, reported that the global AI market is projected to increase by $89.9 billion in 2025. This is more than 20 times the market worth of $4.8 billion in 2017. Thus, it’s not surprising that more and more businesses are investing in AI.
A PwC study revealed that 63% of consumers agree that AI will help solve complex problems that plague our society. Another 50% of consumers believe that AI will provide major solutions for cybersecurity and privacy (68%), cancer and diseases (66%), clean energy (62%), personal financial security and fraud (61%), global education (58%), global health and well-being (56%), economic growth (56%), income inequality (38%), and gender inequality (31%).
For years, AI has been introducing innovations and tools for several purposes. The technology has made it easier than ever to produce images that look completely real but are fake. While this sounds very promising, it can also lead to a new age of mischief and online fakery. Recently, Facebook removed over 900 fake accounts on its platform and Instagram. These fake accounts, which used AI-generated profile pictures, were detected through “inauthentic behavior.”
According to Forbes, a global media company focusing on business, investing, technology, entrepreneurship, leadership, and lifestyle, the fake accounts were associated with a network known as “The Beauty of Life” (or “TheBL”). Results of investigations showed that the global network is linked to The Epoch Times, a controversial conservative news publisher. The network was created mainly to promote an anti-communist, pro-Trump message across hundreds of accounts and pages.
“Our investigation linked this activity to Epoch Media Group, a US-based media organization, and individuals in Vietnam working on its behalf. The BL is now banned from Facebook,” Facebook said in a statement.
Researchers from Graphika and the Atlantic Council’s Digital Forensic Research Lab (DFRLab) recently released a report, showing TheBL used AI to generate fake profile photos. “Dozens of these fake accounts had profile pictures generated by artificial intelligence, in the first large-scale deployment of fake faces known to the authors of this report,” the researchers wrote.
Introducing “Generated Photos”
Most people who have worked for a creative agency or any creative project, in general, have dealt with the world of stock images. While stock images themselves are very convenient and useful, it can be frustrating when you think that the stock image you’re using is also being used by another company. Fortunately, AI has addressed this problem.
Konstantin Zhabinskiy and his team created Generated Photos, a library of faces comprising 100,000 different “people” various expressions, looks, skin colors, hairstyles, and more. Some of you might have probably seen the faces included in the collection. But, the thing is, none of them are real. They were all generated by AI. According to Fast Company, the world's leading progressive business media brand, all of these photos are free to use commercially.
On its webpage, the Generated Photos team said, “These people aren't real! We are building the next generation of media through the power of AI. Copyrights, distribution rights, and infringement claims will soon be things of the past.” They are currently building a richer API, which aims to sort the collection by criteria like age, ethnicity, skin tone, gender, and mood.
It’s also worth noting that the current AI-generated images are much better than they were before. Back in 2014, the faces machine learning techniques generated looked grainy. It would look like a photo taken from a low-quality surveillance camera. However, things have changed. According to Vox, an American news and opinion website owned by Vox Media, today’s images are in full color and are very detailed. Now, the faces resemble people of specific ages and ethnicities.
|Most people who have worked for a creative agency or any creative project, in general, have dealt with the world of stock images / Photo by: gstockstudio via 123RF|
How AI Generated These Images
It’s impressive how AI managed to generate and improve all of the images in Generated Photos. This technique was invented by Ian Goodfellow, the director of machine learning at Apple’s Special Projects Group, back in 2014. The team he was working with spent thousands of hours shooting models in the house. Users don’t have to worry about copyrights for commercial or personal projects because each model signed releases for the database so that you can use the photos freely.
The team was able to get even more photos using the power of AI. As a result, a free and huge database was established for anyone who wants to access it. The main idea of AI-generated images was simple: one algorithm tries to generate a realistic image of an object or a scene, while another algorithm tries to decide whether that image is real or fake. This technique is called “generative adversarial networks” (GANS), one of the most powerful new machine learning technologies.
But aside from photos, combining generative adversarial techniques with other recent advances can also make you a video. Recently, Egor Zakharov and others at the Samsung AI lab in Moscow presented a paper that showed how modern AIs can create a fairly realistic fake video from a single image. The authors demonstrated that it can make fake videos more convincing when it has a few more images to go off.
With the success and continuous growth of AI, it is possible to develop such techniques that can help us in many ways. However, we should still be cautious because AI-generated photos can no doubt be used to manipulate the public.