31/07/2026 19:38 - Tecnologia
The era of disinformation through 'deepfakes' and hyper-realistic fake content might be coming to an end in Europe. According to The Guardian, starting Sunday, August 2, 2026, the European Union (EU) will implement strict rules requiring companies to clearly label any image, audio, or text generated by Artificial Intelligence (AI) that looks authentic.
This historic measure is part of the EU AI Act, a pioneering legislation that aims to establish a clear, risk-based framework for the use of artificial intelligence. While new AI systems entering the European market will need to comply by August 2, existing systems will have four additional months to adapt.
The regulations establish that synthetic content designed to look real must:
Texts on topics of public interest will also need to be labeled if they have not gone through human editorial review.
The rules will not apply to users' personal content or 'evidently artistic', satirical, or fictional works. Additionally, companies are encouraged to label old content, although it is not mandatory.
Non-compliance with these regulations will be costly: fines could reach 15 million euros or 3% of the global turnover of the offending company.
Green MEP Sergey Lagodinsky, who helped negotiate the AI Act, highlighted the importance of this measure: 'This is not just about consumer protection, it is also about protecting democracy. Making this information transparent is something we need to preserve our democracy and the authenticity of facts online.'
In recent years, we have seen how deceptive 'deepfakes' flooded social networks. Fake audios of Slovak opposition leader Michal Šimečka conspiring to manipulate elections went viral, alongside manipulated videos of French President Emmanuel Macron dancing in an 80s disco, and fake images of former British Labour leader Keir Starmer mistreating his staff. The new regulation seeks to put a stop to these harmful practices.
Companies like Google, Meta, and TikTok have already taken proactive measures. Google, for example, implemented its SynthID tool in 2023, managing to place invisible watermarks on more than 100 billion images and the equivalent of 60,000 years of audio. Over 180 organizations have signed a voluntary code of practice promoted by the European Commission to demonstrate their commitment to transparency.
While the tech industry supports the fight against 'deepfakes', some voices, like the Computer and Communications Industry Association (CCIA), fear the regulation might be too broad and saturate users with labels, similar to what happened with cookie warnings. However, the EU tech policy commissioner, Henna Virkkunen, assured that Europe has anticipated this development and created a framework that 'offers legal certainty to innovators while protecting the public interest.'
Definitively, this August 2 marks the beginning of a new digital era, where transparency and trust will be the fundamental pillars of our online information consumption.
Alfredo S. Quiroga
