In the weeks after the September 11 terror attacks, reports surfaced that terrorists might have communicated with each other through messages embedded in images posted on the Web. So far, no such hidden communications have been confirmed publicly, but intelligence agencies are certainly keen on finding them if they exist. To aid in the search, a computer scientist at the State University of New York at Binghamton has developed a way to screen digital images for evidence of hidden content.
Binghamton’s Jessica Fridrich says her algorithms examine the numbers that encode color in pixels, the colored or gray dots that make up an image. When an image conceals information-say, a 15-page text file-the numbers that encode its pixel colors are changed slightly. While the human eye can’t see the resulting color changes, Fridrich’s algorithms can detect statistical anomalies in the underlying numbers. In most kinds of image files, Fridrich’s tool can detect the signatures of a number of concealment-or “steganography”-programs, all widely shared in the hacker subculture. Cryptographers must then decode any images that have been altered.
Fridrich delivered the first version of the software to her U.S. Air Force sponsors last year. “What they do with it, I’m not allowed to know. We can only assume the government is somehow using it,” she says.
“Her work is first rate,” says computer scientist Rafael Alonso, technical director of Web informatics at Princeton, NJ-based Sarnoff. But the software requires investigators to make preliminary guesses about which Web sites might harbor images with hidden messages. In the future, says Alonso, to narrow the range of images to scan, algorithms like Fridrich’s might be combined with search engine software capable of “shining a flashlight in the sewers of the Web”-dredging obscure sites like personal pages and classified ads that are presumably attractive for covert communications.
Fridrich predicts “sharp competition in the next few years” from other approaches to ferreting out hidden messages. Security won’t come easy, though. Information could be stashed in video and music files as well as photos, for example. Still, Fridrich’s tools mark an important first step toward finding pictures that contain thousands of words.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
Data analytics reveal real business value
Sophisticated analytics tools mine insights from data, optimizing operational processes across the enterprise.
Driving companywide efficiencies with AI
Advanced AI and ML capabilities revolutionize how administrative and operations tasks are done.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.