Perhaps unknown to many, Google has been watermarking AI content across audio, video, and images since August 2023 and text since 14th May, 2024 using its Synth ID technology. This is used on all its AI models and AI programs.
While this technology was previously available only to select journalists and media professionals, it has been available to anyone worldwide since October 6th, allowing them to determine whether any images, video, or audio are AI-generated. This will come as a major relief to hapless original content creators, besides helping consumers of content make meaningful and better-informed decisions across the board with regard to its trustworthiness.
In fact, the impact of Google Synth has been humongous already, which can be gauged by the fact that as many as 180 billion images and videos, as well as 240,000 years of audio content, have been watermarked.
For those in the content creation, digital marketing and PR universe, the wide availability of this technology marks a watershed moment in that it will help usher in transparency and authenticity by way of enabling content verification. This was something that was desperately required at a time when AI slop threatened to make content largely untrustworthy and unreliable.
Besides, the EU passed its watermarking law this August, resulting in Generative AI platform Claude starting to embed hidden signatures into AI text. This has made people using AI content quite concerned about its SEO implications.
With fake images, deepfakes, and other kinds of dangers lurking out there for society as a whole because of advances in technology, being able to flag something that is not human-generated is something that was definitely required.
That being stated, it is pertinent to note that content being marked as AI-generated will not have any direct bearing on its ability to rank on Google. The idea behind introducing the tool is not to mete out punishment to content creators for using AI, but to let the consumers of such content be made aware of its source so that they can make informed choices about the information being provided to them.
Its primary usefulness lies in its ability to flag misinformation and misattribution, which are clearly malpractices. Besides, it puts reality back into content in a world where it is increasingly difficult to differentiate between brilliant content created by AI and original content created by human beings themselves.
When you can no longer be sure that what you see, hear, and read has indeed been created by a human or has been brilliantly simulated by AI, the Synth ID technology comes as a huge relief to consumers of content perplexed about its provenance.
With major AI platforms like Anthropic, Nvidia, ElevenLabs, and OpenAI using this technology to identify AI-generated content, it is only natural for content creators and marketers to worry about how the algorithms react to it. According to a test carried out by renowned SEO authority Neil Patel, which involved studying 744 AI-generated articles across 68 websites spanning different industries, articles that comprised human-edited AI text performed a whopping 444 % better than purely AI-generated text, with the results getting even better with time.
This points to an unmistakable truth. While AI-generated content has swamped the market, it offers diminishing yields. Human-generated content may be scarce in comparison, but it garners an outstanding 87% of organic traffic when compared with the piddling 4.9 per cent of organic traffic. The writing is on the wall, and it is time to smell the coffee.