THE CENTER FOR DATA INNOVATION | DATAINNOVATION.

ORG

Why the U.S. Needs a National AI

Strategy
Joshua New
@josh_a_new

What’s Next in Artificial Intelligence

October 30, 2018
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

ABOUT THE CENTER FOR DATA INNOVATION

The Center for Data Innovation is the

leading think tank studying the
intersection of data, technology, and
public policy. With staff in Washington,
D.C., and Brussels, the Center formulates
and promotes pragmatic public policies
designed to maximize the benefits of
data-driven innovation in the public and
private sectors. The Center is a non-profit,
non-partisan research institute affiliated
with the Information Technology and
Innovation Foundation.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

WHY THE U.S. NEEDS A NATIONAL AI STRATEGY

1. Why a strategy is necessary

 Value of AI
 Falling behind the competition
2. What a strategy should look like
 Provide U.S. firms with the resources
necessary to compete effectively;
 Accelerate public sector adoption of AI;
 Use innovation-friendly regulation to
minimize consumer harms;
 Shape trade policies to allow AI to flourish;
 Ensure all Americans can participate in an
AI-driven economy
3. Crafting a National Strategy
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

VALUE OF AI

Future U.S. competitiveness will be

determined by how effectively the public
and private sectors can deploy AI.

AI can provide enormous benefits to

economy, society, and national security.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

FALLING BEHIND THE COMPETITION

The United States has unique

advantages that have made an early
leader in AI, but this lead is slipping.

Other countries are aggressively

pursuing policies to spur AI
development and adoption with
national strategies. The United States
National AI strategies. Source: Politics + AI, Tim Dutton
is not.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

WHAT A STRATEGY SHOULD LOOK LIKE

The ultimate goal of a national AI strategy

should be to make the United States a global
leader in the development and use of AI.

A national strategy should, among other things:

 Provide U.S. firms with the resources necessary to
compete effectively;
 Accelerate public sector adoption of AI;
 Use innovation-friendly regulation to minimize
consumer harms from AI;
 Shape trade policies to allow AI to flourish; and
 Ensure all American workers can participate in an AI-
driven economy.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

PREREQUISITES: DATA

Barriers to data make obtaining

value from AI “theoretically
possible but pragmatically
difficult.”

These barriers are typically

caused by regulatory obstacles or
perverse market incentives to France’s National AI Strategy. Source: Institut
Français 🇪🇪.

collecting, sharing, and using

data.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

PREREQUISITES: PUBLIC R&D

Public R&D is a crucial driver of

innovation, and the U.S. is
dramatically underinvesting.
 Federal R&D intensity has fallen from 1.86
percent in 1964 to 0.62 percent in 2015.
 Federal R&D intensity in basic research
fell to 0.21 percent in 2015, its lowest
since 1980.
 Private R&D is not a substitution for public
Federal R&D intensity 1953-2015. Source: AAAS, R&D.
National Center for Science and Engineering
Statistics.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

PREREQUISITE: HUMAN CAPITAL

The United States already does

not have a workforce with the
skills to develop and use AI
effectively.
• Job postings for “machine learning
engineers” increased by 980 percent
from 2012 to 2017.
• Jobs for workers and managers with
data science and analytics skills will Data scientist. Source: Joonspoon.

increase to 2,720,000 by 2020.

 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

ACCELERATE PUBLIC ADOPTION OF AI

The public and private sectors

face similar challenges with AI,
but the public sector does not
experience pressure to innovate.

Government adoption of AI would

have important secondary
benefits.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

USE INNOVATION-FRIENDLY REGULATION TO MINIMIZE HARM

The United States may lack

appropriate governance that
would maximize the benefits of AI
while minimizing its harms. This
will result in less AI use.

Policymakers should encourage

adherence to the principle of Report: How Policymakers Can Foster Algorithmic Accountability

algorithmic accountability.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

SHAPE TRADE POLICIES TO ALLOW AI TO FLOURISH

The practices of other countries

can have a significant impact on
how effectively U.S. firms can
develop and deploy AI.

All trade negotiations should

consider opportunities to support
Chrystia Freeland, Luis Videgaray Caso, and Rex Tillerson. Source:
U.S. Department of State AI.
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

ENSURE ALL AMERICAN WORKERS CAN PARTICIPATE IN AN AI ECONOMY

Concerns about automation

engender support for
counterproductive policies that
would restrict AI deployment.

It is difficult to predict what kinds

of skills will be necessary to
perform the non-AI work of the Collaborative robot Sawyer on the factory floor at Tennplasco in

future. Lafayette, TN. Source: Rethink Robotics

 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

CRAFTING A NATIONAL AI STRATEGY

There are a variety of different

policymaking mechanisms that
the government could use to
create a national AI strategy:
 Administration-led
FCC’s National Broadband Plan
 Agency-led
 Congress-led
 All of the above
 THE CENTER FOR DATA INNOVATION | DATAINNOVATION.ORG

THANK YOU

 New report: Crafting a National Artificial Intelligence Strategy for

the United States
 Release event on December 4
 How Policymakers Can Foster Algorithmic Accountability:
 http://www.datainnovation.org/2018/05/how-policymakers-can-foster-
algorithmic-accountability/
 Email me: jnew@datainnovation.org
 @josh_a_new