Watermarking for Humans and Machines vs. Algorithmics: Why Fake Images are Not So Popular on Twitter

Water marks work well for large objects like images, songs, or book chapters but don’t work for smaller objects like words or letters. That means these approaches don’t handle content that blends human and machine well. A watermark is a graphical representation of a digital highlighter, and it should be used if you have a document that has been edited by a human.

You can also watermark digital images using steganography, which hides one message inside another cryptographically. First used by spies to smuggle secrets, there are now plenty of design tools that add hidden markings to images, then crawl the web looking for copyright violators. And it works for watermarking too. You can sign a paragraph of text and then tell when it was changed through either a centralized system or a distributed one. This is why the movie you bought only plays in the App Store, and the one you forgot about still belongs to you.

According to experts, the pledges are only half measures. “There’s not going to be a really simple yes or no on whether something is AI-generated or not, even with watermarks Sam Gregory is the program director of Witness, a nonprofit that helps people use technology to promote human rights.

Only a few months ago, AI content was easy to spot: unnatural inflections in speech, weird earlobes in photos, bland language in writing. This is no longer the case. In June, scammers used an AI to impersonate a daughter’s voice and rob her mother. Candidates are using deepfakes. The back-and-forth discussions that are required to separate a mark from money may be automated by the wisdom of LLMs. We need a way to distinguish things made by humans from things made by algorithms, and we need it very soon.

In May, a fake image of an explosion near the Pentagon went viral on Twitter. It was soon followed by images seeming to show explosions near the White House as well. The images appeared to have been generated by artificial intelligence, but not before the stock market began to fall.

Why Ai Is It Needs It Own Alphabet? A Detailed Analysis of Different Approaches to Make an AI Markup Solution

Unicode is the universal numbering system for text, and text is the fundamental building block of the internet. In Unicode, every character has a number. For an example, the Latin capital letter A is a number 41. There are a lot of A’s that are in the same language. There’s Fullwidth Latin Capital Letter A (Ａ, number EF BC A1), Mathematical Bold Capital A (𝐀, number F0 9D 90 80), Mathematical Sans-Serif Capital A (𝖠, F0 9D 96 A0), and plenty of others. The A has a name and a Unicode value, but it may have its own type as well. There is a letter A for Artificial Intelligence.

But these approaches have three fundamental problems. First, coordination is required. By contrast, a good AI markup solution would need to work seamlessly across billions of devices. The markings would have to be copied and pasted into other applications, operating systems or platforms. Any solution that could be accessed to any human with an internet connection would have to be trained immediately. It would need a software update to be deployable.

Fortunately, we have a solution waiting in plain sight. To understand the elegance of this hack, we need to first look at the alternatives and why they will not work.

Source: https://www.wired.com/story/to-watermark-ai-it-needs-its-own-alphabet/

Alistair Croll: Web Performance, Conferences, and Experiments at Strata, Year One Labs/Dark Matter, and Just Evil Enough

Alistair Croll is an author, entrepreneur, and conference organizer. He cofounded the web performance startup Coradiant, the Year One Labs accelerator, and the FWD50 digital government conference. He chaired the world’s leading conference on data science and Strata, and served as a visiting executive at Harvard Business School. Alistair is the author of three books on technology and business, including the best-selling Lean Analytics, and is currently working on Just Evil Enough, a playbook for subversive thinking.