Tim Hwang
Leading U.S. companies are investing in the broad research field of artificial intelligence (AI), but where, specifically, are they making these investments? This data brief provides an analysis of the research papers published by Amazon, Apple, Facebook, Google, IBM, and Microsoft over the past decade to better understand what work their labs are prioritizing, and the degree to which these companies have similar or different research agendas overall. The authors find that major “AI companies” are often focused on very different subfields within AI, and that the private sector may be failing to make research investments consistent with ensuring long-term national competitiveness.Download Full Report
Executive Summary
Within the broad research field of artificial intelligence (AI), it is worth understanding, specifically, what leading U.S. companies invest in. This data brief conducts an analysis of the research papers published by Amazon, Apple, Facebook, Google, IBM, and Microsoft over the past decade to better understand what work their labs are prioritizing, and the degree to which these companies have similar or different research agendas overall.
We find the following:
Major “AI companies” are often focused on very different subfields within AI. While companies like Amazon, Apple, Facebook, Google, IBM, and Microsoft are often grouped together generically as leaders in AI, an analysis of their publications shows considerable differentiation in the areas of research they prioritize. While publications may not provide the full picture of these companies’ research agendas, as companies may not choose to publish on work that will form the basis of valuable intellectual property, it still provides a window into the differences in research agendas between these companies. Policymakers should be careful to consider these differences in framing national assessments of technological competitiveness and in strategizing government investments in research.
The private sector may be failing to make research investments consistent with ensuring long-term national competitiveness. None of the leading companies examined in this analysis appear to be prioritizing work on problem areas within machine learning that will offset the broader structural challenges the United States faces in deploying and benefitting from the technology when competing against authoritarian regimes. This includes work in areas such as few-shot learning, federated learning, simulation learning, interpretability, and ML fairness.
No comments:
Post a Comment