Contents

The Internet Operates Under New Rules Starting Today: A Major Change Has Just Taken Effect

Artificial intelligence (AI) has become an integral part of our daily lives, often operating unseen in the background. However, a significant shift began on August 2nd with the implementation of key transparency provisions within the European Union’s landmark AI Act. These new regulations are designed to bring AI’s presence into the forefront, ensuring users are aware of when they are interacting with or consuming AI-generated content. Companies are now explicitly required to label AI interactions, disclose algorithmic emotion analysis, and clearly mark AI-generated visuals or audio online.

The Digital Landscape is Evolving: A New Era of AI Transparency

The new transparency rules, effective from August 2nd, mark a pivotal moment in the governance of artificial intelligence. These regulations introduce several critical obligations:

• Disclosure of AI Interaction: Users must be informed when they are engaging with an AI system.
• Labeling of Biometric Systems: Systems using emotion recognition and biometric categorization must be clearly identified.
• Identification of AI-Generated Content: Deepfakes and other AI-generated media must be explicitly disclosed.

This mandate specifically targets graphics, video, and audio utilized in professional, commercial, or public capacities. It’s a proactive step to address concerns about AI-generated fake news and disinformation, providing users with essential context.

Who Do These Regulations Affect?

The scope of the AI Act extends beyond just technology providers. It impacts a broad spectrum of businesses that leverage AI in their interactions with people, including but not limited to:

• Banks and financial institutions
• Insurance providers
• E-commerce platforms and online marketplaces
• Media organizations and content creators
• Marketing agencies
• Producers of audiovisual content

It’s important to note that strictly personal, non-professional applications of AI are generally exempt from these transparency requirements.

What These New AI Rules Mean in Practice

Direct Interaction with AI Systems

The AI Act mandates clear communication whenever a user encounters an AI system. This applies to a wide range of interfaces, such as:

• Chatbots embedded on websites
• Voice assistants and voice bots on customer service hotlines
• Virtual advisors within mobile applications

If, until now, you were unaware that you were conversing with an AI, companies are now legally obligated to make this fact known. This brings much-needed clarity to digital interactions, as bots and AI now generate significantly more internet traffic than humans.

Emotion Recognition and Biometric Categorization

Another crucial aspect addresses the use of emotion recognition and biometric categorization systems. When AI analyzes facial expressions, vocal tone, or other unique facial features, this must be explicitly disclosed to the individuals concerned. A company employing such tools can no longer “silently” monitor a user’s reaction to an advertisement, conversation, or video. They must provide clear notification that an algorithm designed to interpret emotions is actively at work.

Addressing Deepfakes and AI-Generated Content

Mandatory Labeling of Deepfakes

The AI Act specifically requires “deepfakes”—images, sounds, and videos generated or manipulated by AI to appear authentic—to be clearly labeled as artificial content. On a technical level, providers are obligated to implement machine-readable identifiers (metadata) for AI-generated content across various formats, including text, images, audio, and video. This metadata allows systems to automatically detect and identify synthesized media, promoting transparency and combating misinformation.

Nuances for Artistic and Informational AI Content

For artistic, satirical, and fictional content, the labeling requirement is somewhat relaxed. Information about AI use might appear, for example, in the end credits of a film, to avoid disrupting the immersive experience. However, advertisements, as a general rule, should be labeled at their initial exposure without such allowances.

AI-generated texts are not categorized as deepfakes, but they carry their own transparency obligations, particularly when they aim to inform the public about matters of public interest. This explicit obligation typically does not apply to informational texts that have undergone human verification and for which editorial responsibility lies with a natural or legal person, ensuring a human oversight layer for critical information.

Bringing AI Usage into the Light: What to Expect

In practice, these regulations mean that internet users will increasingly encounter “AI” or “synthetic content” labels, or hear voice announcements indicating interaction with an AI system. This will be evident whether you’re chatting with a customer service bot or viewing potentially manipulated video footage online.

While these new provisions don’t guarantee that all content will be inherently trustworthy or safe, they empower users with the crucial awareness that they are engaging with AI-generated material. The mandatory labeling ensures a baseline level of transparency. Furthermore, the European Commission’s AI Office has been granted new powers to enforce these regulations on providers. Non-compliance can lead to substantial penalties, including fines of up to €15 million or 3% of a company’s global annual turnover, underscoring the serious commitment to AI accountability.

Source: Gemini

Opening photo: Gemini

About Post Author

Deepak Malik

See author's posts