A recent study by the Meta Oversight Board found that major AI systems, including those built in the U.S., are more likely to refuse to criticize restrictive leaders or governments. This raises concerns that the large language models powering chatbots and AI agents could be regurgitating and spreading government influence over online speech as the technology is increasingly adopted worldwide.

Study Findings

The study picked 10 commercial large language models by top tech companies and asked the AI systems to make critical pamphlets, write limericks, give reasons if someone should join protests, and more. The results showed that AI models are reflecting speech restrictions beyond the countries where they apply.

For example, when asked to create a pamphlet critical of President Donald Trump or Britain’s King Charles III, the AI chatbot obliged. However, when asked to do the same for Thailand’s king, Saudi Arabia’s crown prince, or China’s leader, the artificial intelligence model declined.

Implications

The study indicates that AI models are likely not helping potential demonstrators in free countries create protest materials to speak out against events in restrictive countries. This has the practical effect of extending the long arm of restrictive governments across borders to limit speech in free countries.

The board said it could not determine the causes for the responses but suggested that models could have absorbed latent biases in data used to train the systems and companies might have weighed the risks and liabilities.

Original reporting: Dallas TX News (HLL/CB) — read the source article.