Critics fear that the revised fake law has transformed from targeted AI deepfake protection into a radical censorship power.
According to digital rights advocates, what began as a seemingly rational attempt to tackle AI-generated deepfakes has snowballed into something much more troubling. Many controversial nurturing Originals, Foster Art, and Keep Entertainment Safe (fake) behaviour are originally intended to prevent people from illicit digital replicas, but they threaten to fundamentally change the functionality of the Internet.
The bill’s expansion has led to alarm bells ringing across the tech community. It could create a drastic censorship framework, not just protecting celebrities from fake videos.
From wise safeguards to sledgehammer approaches
The initial idea was not completely misguided. Creating protection against AI systems is to generate fake videos of real people without permission. We’ve all seen the uneasy deepfakes circulating online.
But rather than creating narrowly targeted measures, lawmakers chose what the Electronic Frontier Foundation calls a “federalized image licensing system” that goes far beyond reasonable protection.
“The updated bill will double with that first, false approach,” EFF said.
What’s particularly bothering is the No Fakes Act requirement on almost all internet platforms to implement a system that removes content, not only after receiving the Takedown notification, but also prevents similar content from being uploaded again. Essentially, it forces the platform to deploy content filters that have proven unreliable in other contexts.
The calm of innovation
Perhaps the most concern for the AI sector is how no-fax acts target the tool itself. The revised bill doesn’t just chase harmful content. It could shut down the entire development platform and the entire software tool that can be used to create malformed images.
I feel this approach is reminiscent of trying to ban word processors as someone might use it to write defamatory content. The bill includes several restrictions (for example, tools must be “mainly designed” to create fraudulent replicas, or for limited other commercial uses), but these distinctions are subject to interpretation.
Small UK startups challenging AI image generation can find themselves caught up in an expensive legal battle, based on flimsy allegations, long before they have the opportunity to establish themselves. Meanwhile, tech giants with the armed forces of lawyers can better survive such a storm and potentially entrench their advantage.
Anyone who has worked on YouTube’s ContentID system or similar copyright filtering tools knows how frustrating it is. These systems routinely flag legitimate content such as musicians who use materials to play their own songs and creators under fair trading regulations.
The No Fakes Act effectively mandates similar filtering systems across the Internet. It includes parody, satire and commentary carve-outs, but it has proven practically impossible to implement these distinctions algorithmically.
“These systems often flag similar but not identical,” explains EFF.
For small platforms that do not have Google scale resources, implementing such filters can prove to be extremely expensive. Possible outcome? Many simply overcensor the censorship to avoid legal risks.
In fact, you might expect big tech companies to oppose such drastic regulations. However, many remain prominently quiet. Some industry observers suggest that this is not a coincidence. Established giants can more easily absorb the compliance costs of crushing smaller competitors.
“It’s probably no coincidence that some of these very giants are OK with this new version of fake,” EFF points out.
This pattern repeats throughout the history of technical regulations. What appears to be the regulations governed by big technology often solidifies the marketplace by creating barriers that are too costly for newcomers to overcome.
Fake acts do not threaten anonymous speech
What is listed in the law is another troubling provision that could make anonymous internet users public based on mere claims. The bill allows anyone to obtain subpoena from the court clerk without judicial review or evidence – enforces them to identify information about users accused of creating fraudulent replicas.
History shows that such mechanisms are ripe for abuse. Critics with valid points may be unmasked and their commentary may potentially be harassed if they include screenshots and quotes from people trying to silence them.
This vulnerability can have a significant impact on legitimate criticism and whistleblowing. Imagine exposing corporate fraud only to reveal your identity through the rubber stamp subpoena process.
Given that Congress recently passed the Take It Down Act, this push for additional regulations seems strange. The law itself has raised privacy concerns, particularly focusing on monitoring encrypted communications.
Rather than assessing the impact of existing laws, lawmakers seem determined to move forward with wider restrictions that could restructure internet governance over the coming decades.
As the No Fakes Act moves through the legislative process, the coming weeks will prove to be important. For those who value internet freedom, innovation and a balanced approach to the challenges of new technology, this is what you’re actually watching.
(Photo: Markus Spike)
reference: Openai File: Betray the safety of ex-staff billing profit greedy AI
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London. The comprehensive event will be held in collaboration with other major events, including the Intelligent Automation Conference, Blockx, Digital Transformation Week, and Cyber Security & Cloud Expo.
Check out other upcoming Enterprise Technology events and webinars with TechForge here.
