Home > News

AI tech designed to end video game leaks for good uses watermarks hidden “in plain sight”

An AI-powered way to discourage game devs from sharing leaks
Last Updated on
AI tech designed to end video game leaks for good uses watermarks hidden “in plain sight”
PC Guide is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Read More

New AI watermarking technology designed to prevent video game leaks is currently being developed. Previewed at GDC earlier this month, the tech allows video game developers and publishers to protect their biggest secrets “in plain sight” by using a technique known as steganography. The gaming industry often suffers from leaked content or footage being released ahead of an anticipated launch – personally, the biggest that comes to mind is the Half-Life 2 leak way back in 2003.

Keeping early previews and information under wraps is a constant challenge for developers. CEO and Co-founder of EchoMark, Troy Batterberry, led a presentation titled “Stopping Insider Leaks Through the Use of AI-powered Steganography” during GDC, presenting developers with a way to protect their IPs and other sensitive information.

Using AI to identify how information is being leaked

Whatever your stance on the involvement of AI in gaming, using artificial intelligence to prevent game leaks seems like a massive time saver that creates peace of mind for developers. At this year’s GDC, Batterberry called the tech “a whole other level of watermarking,” according to Game File (who was in attendance).

Steganography refers to the process of seamlessly hiding information within other data to conceal the original content. EchoMark uses AI to hide unique identification marks with this method – one of the primary tools can make thousands of small adjustments to the spacing of lines within documents – something a human won’t notice, but artificial intelligence will.

This means copies of documentation shared across emails are ever-so slightly different, and if something is leaked, the AI will be able to identify which employee’s copy was shared. It’s referred to as having a unique “fingerprint” for every recipient of sensitive information. It may also rephrase words or slightly change the structure of sentences, resulting in possibly over a trillion different versions, which can be identified by AI.

In addition, EchoMark’s CEO demoed a new “alpha blend” technology that overlays a semi-transparent watermark onto computer screens. It’s practically unnoticeable for a human to see, but a neural network can trace any screenshots or photos back to the culprit if something leaks out. Essentially, it’s an in-house piece of surveillance that should help discourage information leaks, especially when there are legal consequences.

Game File claims organizations are already implementing this without notifying employees, though in the case of movie studios, ensuring individuals know their copies are individualized is a more standard practice to discourage “slippery” behaviour. In short, it should cut ties to reliable, anonymous insider leaks and leave those responsible in a heap of trouble.


1987
DLSS 5 first impressions

Are you interested in using DLSS 5 with neural rendering enabled?

About the Author

At PC Guide, Jack is mostly responsible for reporting on hardware deals. He also specializes in monitors, TVs, and headsets and can be found putting his findings together in a review or best-of guide.