Rapid advances in artificial intelligence are increasingly creating immediate societal disruptions, ranging from corporate layoffs to mounting environmental costs. The expansion of data centers has triggered sharp increases in energy consumption and carbon emissions, while malicious actors are utilizing AI agents to automate disinformation campaigns. Additionally, algorithmic biases continue to cause discriminatory outcomes, and military applications of AI for target identification are leading to greater loss of life.
Despite these pressing issues, global policy discussions are frequently dominated by apocalyptic warnings regarding existential risks. Prominent figures and decision-makers have amplified these concerns, including United States Senator Bernie Sanders, who has championed bans on superintelligence, and former Secretary of State Hillary Clinton, who described open-source AI as a genuine threat. Cautionary warnings have also emerged from the Vatican.
Critics argue that intense focus on hypothetical doomsday scenarios is being used to justify restrictive regulations that target open-source AI models. Because building frontier AI models from scratch can cost upwards of $100 million in compute power alone, developing nations and the Global South remain heavily dependent on technology built by a small number of Western tech giants. Opponents of strict open-source restrictions compare this dynamic to historical colonial practices, warning that shutting down free or lower-cost customizable models locks users into severe data extraction, cognitive dependency, and absolute reliance on foreign corporate monopolies.
The risks of relying on imported models were recently demonstrated in Kenya during the rollout of its AI-driven Social Health Authority (SHA) system. Built on assumptions that failed to account for the country's extensive informal economy, the imported system failed by systematically overestimating citizen incomes and overcharging the poorest residents for healthcare premiums.
English (US) ·