Wahington: The White House announced on Friday that some of the top names in tech, including Google, Meta, OpenAI, and others, have promised to take steps to make AI-generated content safer and more transparent. One of the measures they have agreed to implement is adding watermarks to AI-generated text, video, audio, and images.
AI-generated content refers to digital material that is created or manipulated by artificial intelligence tools, such as deepfakes, synthetic voices, image generators, and text synthesizers. While these tools can have positive and creative applications, they can also pose risks of spreading misinformation, deception, fraud, and harassment.
According to the White House statement, the tech companies have made voluntary commitments to develop robust mechanisms, such as provenance and watermarking systems, for audiovisual content that is not readily distinguishable from reality or that is designed to deceive or mislead users. The watermarks would indicate which AI system was used to create the content, but not reveal any personal information about the user.
The statement also said that the companies have pledged to test their systems for safety and fairness, share information and best practices with each other and the public, and respect user privacy and consent. The White House said that these commitments are aligned with its efforts to promote trustworthy AI and protect democratic values.
The move comes amid growing concerns about the potential harms of AI-generated content, especially in the context of social media platforms and online communities. Earlier this year, a fake image of former President Donald Trump’s arrest went viral on Twitter, which was later revealed to be created using an image generator called Midjourney. In another case, a political ad supporting Florida Governor Ron DeSantis used an AI-synthesized voice of Trump to endorse him.
The White House said that adding watermarks to AI-generated content would enable creativity with AI to flourish and reduce the dangers of fraud and deception. It also said that it would continue to work with the tech industry, civil society, and international partners to advance the responsible and ethical use of AI.
