Cartoon: America’s AI-safety debate

The debate regarding AI safety is intensifying as advanced systems evolve. Daniel Kokotajlo, a former OpenAI researcher and author of the AI 2027 forecast, warns that misaligned AI could potentially threaten humanity. He argues that this risk does not stem from malicious intent, but rather from the possibility that sufficiently capable systems could acquire significant hard power, leading them to ignore human constraints or stop following intended directions, thereby creating a systemic risk to human existence.