India Proposes Strict Regulations for AI Content
India’s Ministry of Electronics and Information Technology (MEITY) has put forth a set of stringent regulations regarding content produced by artificial intelligence (AI). This proposal calls for clear labeling, metadata traceability, and transparency for all public media created by AI.
The regulations stipulate that AI-generated content must be distinctly marked, with the label occupying at least 10 percent of the visible area in visual media or 10 percent of the audio clip’s length.
Social media platforms will be required to obtain explicit confirmations from users regarding AI-generated content uploads and are urged to strongly verify these claims. Furthermore, these platforms will have broader authority to label and remove content believed to be AI-generated without the correct declarations.
The MEITY’s proposed rules mark one of the widest definitions of AI content so far, encompassing any information that is “artificially or algorithmically created, generated, altered, or modified using computer resources in a manner reasonably thought to be authentic or truthful.”
According to the ministry, recent occurrences of deepfake audio and video being spread on social media highlight the potential for generative AI to craft credible falsehoods, misrepresenting actions and statements that individuals did not actually perform.
This kind of content could be misused to propagate misinformation, harm reputations, influence elections, or carry out financial scams, the ministry warned.
Ashwini Vaishnau, Minister of Information Technology, noted that the proposed amendment to India’s Information Technology Regulations aims to enhance accountability for both internet companies and individual users.
Vaishnau mentioned that discussions are already underway with major AI firms, which have indicated that metadata can help identify content altered by AI. While initial responsibility for labeling AI-generated content lies with these companies and their users, social media networks bear the ultimate duty to ensure proper labeling aligns with community guidelines.
The concerns about realistic and easily created AI deepfakes transcend mere social media issues, as highlighted by the MEITY statement.
A report by Gartner, a cybersecurity consulting firm, released in September reveals that 62% of organizations surveyed have faced cyberattacks utilizing AI deepfake technology. This technology is especially advantageous for hackers employing social engineering tactics to trick individuals into revealing passwords or other sensitive information.
Notably, many recent cyberattacks have involved social engineering, where hackers send convincing emails from trusted sources to extract information or install malware. AI not only enhances the realism of these phishing emails but also enables large-scale generation, increasing the chances of victim engagement.
Gartner’s findings also showed that 32% of participants reported facing attacks where hackers deceived AI systems through clever phrasing to make those systems perform unwanted actions.
Gartner analyst Akif Khan warns that as GenAI attacks related to phishing, deepfakes, and social engineering become more widespread, other threats, such as attacks on GenAI application frameworks, are also on the rise.
In a related matter, a major Indian entertainment company has filed several complaints against parties using AI to replicate and distort intellectual property. Notable actors Abhishek Bachchan and Aishwarya Rai have initiated lawsuits against Google and YouTube, claiming that AI-generated content falsely depicting them in “fictional” and “explicitly sexual” scenarios. Their lawsuit emphasizes that by allowing deepfake content to circulate, social media platforms are inadvertently training AI in a way that potentially misrepresents their image.
The concern is that as misinformation spreads through countless user accounts, the misleading representations of celebrities like Bachchan and Rai can overshadow their genuine work.
