How government can light up AI
- By Matt Leonard
- Oct 09, 2018
What: "Reducing Entry Barriers in the Development and Applications of AI," a new report from R Street Institute, a libertarian think tank
Why: Facilitating artificial intelligence development can help address concerns about lagging productivity and the slow spread of AI across the economy.
Findings: With future economic stability likely depending on AI, the U.S. must focus on boosting the number of skilled AI workers, support the development of high-quality datasets and specialized hardware and reduce barriers to entry for new AI firms.
The strong demand for AI specialists benefits large, established firms that can afford to pay the high salaries these professionals command. To level the playing field, the government should reform the immigration system to allow more highly skilled workers into the country and allow companies to deduct the costs to train AI analysts.
To increase the number of datasets on which to train AI algorithms, the report suggests the government release more of its data in machine-readable form and clarify the fair-use exemption for training data. Data that cities and municipalities might also have on traffic patterns or electricity could also be of interest to AI trainers and developers.
Because AI development tends to require specialized hardware, the government should ensure that the supply chain for these products remains stable so smaller companies are not priced out of the market. Similarly, it should nurture an open ecosystem that includes cloud services like Amazon Web Services and Google Cloud that offer high-performance AI processors and chips as part of their platforms.
Finally, if government reduces identified barriers to entry for AI entrepreneurs, the rate of innovation will increase, allowing AI to spread more quickly across the economy, spurring productivity gains and making the U.S. more competitive internationally.
Read the full report here.
Matt Leonard is a former reporter for GCN.