The author argues that current AI future predictions are undermined by implicit assumptions about the nature of future agents, whether they be humans, institutions, or AI models. These ontological commitments shape theories of risk and governance, leading to divergent forecasts. The piece suggests that while standardization and predictability are desirable for governance, the inherent complexity and potential for unforeseen discontinuities, like novel drone warfare in Ukraine, challenge linear forecasting models. The author posits that our current societal metrics are human-centric and may be inadequate for a future with new types of agents. AI
IMPACT Challenges current AI forecasting methods, suggesting a need for new frameworks to account for agent evolution and unpredictable discontinuities.
RANK_REASON The item is an opinion piece discussing assumptions in AI forecasting.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
