Zuckerberg defends AI as EU rules push watermarking

Mark Zuckerberg has published a 6,500 word defense of AI, positioning himself as a tech optimist at a moment of rising scrutiny for the industry. He argues that AI will not take away people’s jobs, should be open sourced, should avoid restrictive guardrails, and will enrich rather than cheapen human life.

The message follows earlier manifestos from Netscape co-founder Mark Andreessen in 2023 and OpenAI’s Sam Altman in 2024. Zuckerberg’s intervention comes as AI companies face criticism over large investments, limited profits, calls for regulation, backlash over layoffs and data centers, concerns about rogue systems, and unease around products such as Meta Glasses.

The European Union’s Artificial Intelligence Act, which came into force on 2 August, is already changing company practices. Under Article 50, providers must disclose AI-generated images, text, audio or video, and chatbots must make clear that users are interacting with AI.

Anthropic said content generated by its AI models will be watermarked at the model level, allowing the marker to travel with copied and pasted text and persist through some edits. Google also said visible watermarks can be removed from Gemini-generated content, while invisible SynthID watermarks and C2PA metadata will remain embedded.