Freedom on the Net 2023: The Repressive Power of Artificial Intelligence

For the thirteenth year in a row, there has been a drop in internet freedom worldwide, with digital repression causing the largest decline in Iran. Myanmar was found to have the worst internet freedom conditions in the world, while President Rodrigo Duterte's use of an antiterrorism statute to restrict news sites critical of his administration made matters worse in the Philippines. After a presidential candidate whose campaign manager employed internet trolls to intimidate media outlets was elected, Costa Rica's reputation as a champion of internet freedom came under danger. Attacks on the right to free speech have become more widespread; out of the 70 nations that Freedom on the Net covers, 55 have reported facing legal consequences for online speech, and 41 have executed or killed individuals for their comments posted online.

With 47 governments using commenters to sway online debates, generative artificial intelligence (AI) poses a serious challenge to online disinformation tactics. Disinformation strategies have intensified as a result of the increased sophistication, accessibility, and ease of use of AI-based technologies. Governments have also improved and honed their online censorship strategies; in 22 countries, laws have been passed requiring or rewarding digital companies to use machine learning to filter out objectionable social, political, and religious content.

The defenders of democracy must apply the lessons they have learnt from previous internet governance issues to AI to preserve online freedom. AI has the potential to be a powerful tool for digital repression, increasing the efficiency, speed, cost-effectiveness, and ease of censorship, surveillance, and the production and dissemination of false information.