The right to reliable information: the gaping hole in India’s AI Impact Summit declaration

By hosting the summit, India hoped to establish itself as a key player in international AI governance. The customary declaration published at the end of the summit, on 21 February, was meant to conclude the week of diplomatic talks, like in Paris in 2025. The end result is astonishingly disappointing.

The textnotably calls for "international cooperation," the development of trust in AI, and "human capital," and improved energy efficiency in AI development. However, it fails to make a single formal commitment and addresses none of the problems already associated with AI. Major known issues related to the impact of AI, such as information integrity, deepfakes, the fight against disinformation and media sustainability, are completely absent from the text.

Cooperating without solid regulation

The Paris Declaration for an Inclusive and Sustainable Artificial Intelligence for People and the Planet, published in February 2025, at least had the virtue of explicitly mentioning "information integrity." The New Delhi Declaration broke with this, opting to wax lyrical about the benefits of AI for "social good" instead. To ensure the development of trustworthy AI, the text calls for "industry-led voluntary measures" and for "appropriate policy frameworks that enable innovation while promoting the public interest," without clearly defining that public interest. Yet at the European level, RSF has witnessed the failure of self-regulation in the industrywhen it comes to defending the right to reliable information.

With 89 signatories, diplomatically, the New Delhi Declaration looks like a success; issued a year earlier, the Paris Declaration only obtained 64. But this success is misleading. The New Delhi Declaration was co-signed by Donald Trump's United States, which notably boycotted the Paris Declaration. Before signing in New Delhi, the United States delivered a clear message at the summit: it rejected any "global governance of AI" and any risk-based regulatory approach, which it described as "ideological" and synonymous with "bureaucracy." Instead, it called for national frameworks that respect intellectual property- among other things - and invited "nations to join the U.S. as partners to build the AI future." In other words: cooperation that does not involve international regulation of AI.

Putting the right to reliable information at the heart of AI governance

RSF calls for an international AI governance that takes the right to reliable information into account and involves representatives of the media sector, in line with Article 10 of the Paris Charter on AI and journalism, an RSF initiative published in November 2023.

The information risks linked to AI - deepfakes targeting journalists, the corruption of chatbots by Russianand Chinesepropaganda and the industrial production of fake news sites generated by AI- are real and have been clearly documented by RSF. Given these very serious threats to information integrity, the time for statements of intent is over. It is time to act, time to adopt clear, enforceable and binding rules.