AI-Involved US Strike Hits Civilian Area, Threatening Democracy

An incident has occurred where a school was hit in a US airstrike involving AI in command, killing over 170 civilians. In the US airstrikes carried out in Iran, AI was involved in tasks ranging from target identification and providing coordinates to prioritizing efficient strikes. However, information that part of an Iranian naval base had been converted into a school was not incorporated into the AI's decision-making. The US military is reported to have neglected to verify the AI's judgment, failing to prevent civilian casualties. This incident highlights the risks of losing control and insufficient verification in AI's processes for target identification and strike prioritization.

Last month, the CEO of Google DeepMind warned of the dangers posed by rapid technological advancements at the 'AI Impact Summit' held in India. He highlighted the challenges of controlling technology and the potential for losing system control. Historian Yuval Harari has also defined AI as an agent capable of independent decision-making, cautioning against the side effects and potential misuse of the technological revolution. He views the potential misuse of AI as potentially more destructive than nuclear weapons and presented a critical perspective on situations where AI could dominate the world militarily and economically.

Concerns are also rising about the deepening bias in public discourse as reliance on AI recommendations and decisions increases. A 2022 US study found that bots, despite accounting for only 5% of X (formerly Twitter) users, generated 21-29% of all content. This has led to critiques that an environment shaped by AI-driven content decisions can undermine the rational decision-making structures of democracy.