AI development is on a trajectory to automate many jobs, and Americans are worried. 71% of Americans are concerned and 63% support government action, which makes sense given OpenAI's aim to create AI that outperforms humans in most jobs.
AI’s capabilities are growing at a remarkable pace, and how jobs are affected will change with continued development. AI is now outperforming humans across many different tests. General-purpose AIs, AIs capable of functioning across modalities and different types of tasks, are further broadening the range of tasks that could be affected.
Change is coming. While we can’t be sure of the exact effects AI will have on the workforce, research agrees that change is coming, and that it might come fast. One report found that AI might automate 50% of the average workday by 2045. Specific projections differ across studies, but there is some agreement on what and who is at risk:
Large-scale unemployment is possible. AI is only improving as further money is poured into it, and will likely eventually be capable of full automation, of automating all human work. We’ve never before seen a technology that surpasses humans in all tasks, and it would break the previous trend of us adapting from automation by creating new work for ourselves, as AI will be capable of doing the new work too, at a much cheaper rate. The average AI researcher estimates a 50% chance of AI automating all human jobs before 2116.
Legislation can address these risks:
Read the full report here.
Inspecting the claim that AI safety and US primacy are direct trade-offs
Policymakers and engineers should prioritize alignment innovation as AI rapidly develops
The rapid growth of AI creates areas of concern in the field of data privacy, particularly for healthcare data