- Watermarking allows developers to control the watermark encryption, ensuring better content detection and authenticity of AI-generated content.
- AI content detection will foster trust in India’s AI ecosystem and promote the safe adoption of AI.
New Delhi, 19 September 2024: As AI technologies become more advanced, concerns are growing about the source and authenticity of digital content. The latest EY-FICCI report, ‘Identifying AI Generated Content in the Digital Age: The Role of Watermarking’, addresses the critical need to distinguish AI-generated content from human-generated material. It highlights the importance of enabling consumers to identify AI-generated content, ensuring informed decision-making in the digital landscape.
AI content detection tools are essential for establishing authenticity and maintaining the integrity of AI-generated content. As AI evolves, it becomes increasingly challenging to differentiate between human-created and machine-generated content, leading to potential issues such as misinformation, copyright infringement, and a loss of credibility in digital content.
Sharing his insights, Rajnish Gupta, Partner, Tax and Economic Policy Group, EY India, said “Watermarking is a key solution that enables developers to encrypt watermarks, providing better detection and authenticity of AI-generated content. Robust watermarking must resist tampering, and detection systems should maintain low false-positive rates while functioning across different Gen AI platforms. India aims to lead in this area by promoting the development and adoption of advanced watermarking techniques to ensure secure and authentic AI content creation.”
"As generative AI reshapes the digital landscape, we as a nation must come together and establish robust safeguards to trace content origins, ensuring transparency and trust in the AI-driven world. Let us lead with responsibility for a secure and innovative future.", said Jyoti Vij, Director General, FICCI
Key issues in AI generated content
As AI technologies evolve, generating text, images, videos, and audio that are often indistinguishable from those created by humans, there is a growing concern about the authenticity and source of the content which may lead to deepfakes, copyright infringements, fake news, social manipulation, and false attribution and lack of content detection. AI-generated content is now abundant and often indistinguishable from human-created content, significantly influencing public decisions and opinions.
Building trust on AI through content detection mechanisms
One approach to establish authenticity and build a robust AI content detection mechanism is watermarking. Implementing watermarking solutions will be the key to ensuring trust in the future digital economy and the broader AI landscape. Watermarking will enable an ecosystem of accountability, transparency and reliability to foster trust in AI systems.
Governments worldwide are recognising the potential of watermarking technologies. Taking a lead, India can play a pivotal role in shaping a robust domestic digital ecosystem that is secure, transparent and trustworthy.
Need for multi-layered approaches, encompassing both technological advancements and frameworks
Global leaders have also acknowledged the need for frameworks and initiated watermarking and other content detection measures. Policies are being formulated to promote the adoption of watermarking and establishing global standards for its implementation. There is a push to develop open-source tools and digital public infrastructure to support watermarking initiatives. Private sector entities are encouraged to rigorously assess their AI usage, conduct thorough internal testing of AI watermarking, and invest in its widespread implementation. India and the global community need to strengthen frameworks, ensuring legal recognition and governance for digital watermarks across jurisdictions. Ethical implications should also be considered, striking a balance between protecting content and respecting individual privacy. Addressing these challenges is crucial for making watermarking a reliable and widely accepted tool for tracing and verifying AI-generated content.