Real harms from AI today

AI is already causing documented damage. Companies are laying off workers, data centres spike energy consumption and carbon emissions, malicious actors use AI to spread disinformation, algorithms show bias, and AI identifies targets in warfare.

Apocalyptic warnings dominate debate

Instead of addressing these immediate crises, global discourse focuses on existential AI risks. Senator Bernie Sanders champions superintelligence bans, the Vatican issues warnings, and Hillary Clinton calls open-source AI a real threat.

Fear justifies corporate control

Tech giants use doomsday narratives to justify regulations targeting open-source AI models. The logic is simple: if AI is a potential weapon of mass destruction, only trusted corporations should develop it.

“if advanced AI is a potential weapon of mass destruction, it can only be developed by a trusted few”

Digital colonisation mirrors history

This strategy mirrors colonial tactics: preventing local industry from developing to force reliance on the empire's exports. Building frontier AI systems costs upwards of one hundred million dollars in compute power alone.

Kenya's AI system failed locally

Kenya's Social Health Authority system, built on imported AI models, failed spectacularly because it did not capture the reality of Kenya's vast informal economy, systematically overestimating incomes and overcharging the poorest citizens.

Open-source AI bridges the gap

Open-source AI allows local innovators to adapt frontier systems to local data. Developing nations need AI for energy, healthcare, agriculture, manufacturing and climate mitigation, not just conversational chatbots.

Monopolies create dependency

AI monopolies lock users into data extraction and cognitive dependency. Users increasingly outsource critical thinking and creativity to frontier AI systems, creating a feedback loop that centralises control and strips away human agency.

Bans would abandon progress

Abandoning AI would mean losing a technology positioned to accelerate drug discovery and solutions to energy and climate crises. AlphaFold and MOFSeq already demonstrate this potential.

Decentralise, don't consolidate

Safeguarding AI's future requires decentralised, transparent, globally distributed development. Fund robust safety research and rigorous testing, but do not allow fear of hypothetical futures to justify consolidating power in the present.

“Safety does not negate equitable access to AI”