Experts Divided on Timeline for Achieving Human-Level AI
-
There is debate over when AI will reach human-level intelligence (AGI), with some AI leaders predicting it within years while experts and forecasters think decades.
-
The "scaling hypothesis" suggests that continuing to scale up AI systems with more data and compute will inevitably lead to AGI.
-
Surveys show AI experts think there's a 50% chance of high-level AI by 2047, while "superforecasters" think more like 2070.
-
There are high stakes to predicting AI progress accurately due to concerns like deepfakes and bioweapons.
-
Some experts argue policymakers should prepare safety measures now in case powerful AI comes sooner than expected.