.svg)
The Rise of AI In the Federal Workforce
In our current moment, there is unprecedented passion and heat around artificial intelligence (AI) and machine learning (ML). With all the attention our political leaders are devoting to the topic, it begs the question: How is the U.S. federal government, the largest employer on the planet, utilizing these new technologies?
One interesting measure? Federal job postings.
At Citizen Codex, we analyzed publicly available data from USAJobs on more than four thousand postings over the past five years related to artificial intelligence, data science, and machine learning. What our findings reveal is the rapid, upward growth of AI and ML job openings in the federal government.
While there are seasonal patterns to when these jobs are posted, the growth in postings over time is consistent across many different types of agencies. Aside from the traditional defense agencies, the Centers for Disease Control, the Veterans Health Administration, and the U.S. Geological Survey all emerged as leaders in the race for AI/ML talent.
But this is only part of the story.
Large swaths of the defense, security, and intelligence communities do not utilize USAJobs (e.g., the CIA and FBI). Furthermore, for every civilian employee, there are roughly two to three contractors employed by the federal government—roles that are also not captured in USAJobs. And while the effect of artificial intelligence on the federal contractor employment base remains unknown, many of the publicly available AI use cases described at AI.gov were built by contractors.
Even with these contractors not being captured by the data, there was still a 46% annual growth rate in AI/ML federal job postings. The surge in demand for AI/ML hiring is clear. There are several possible catalysts for this:
- Law and Policy. Congress’s passage of the Foundations for Evidence-Based Policymaking Act of 2018 motivated significant investment in using data for decision-making. Executive Orders 13859 and 13960 likely spurred further action, working respectively to establish a comprehensive national strategy for AI and creating the National Artificial Intelligence Initiative.
- Leadership. Many departments and agencies created Chief Data Officer (CDO) roles. This empowered key decision makers to think strategically about the future of AI in government. In January 2021, the U.S. Office of Personnel Management and the Federal Chief Data Officers Council put out a government-wide consolidated data science job application that received over 500 applications in the first 48 hours.
- Crisis. The COVID crisis spurred a data and analytics frenzy. Leaders were hungry to understand not just the state of the disease, but the influence of the pandemic on different aspects of the economy, the country, and the world. This likely explains why the CDC had the highest number of job postings across all non-defense agencies.
Then of course there’s ChatGPT. Most of the activity observed in USAJobs occurred before OpenAI's announcement of ChatGPT in November 2022. ChatGPT, and interest in large language models and AI, will likely increase demand for use of AI/ML models. These technologies will generate a whole new class of federal jobs, such as AI auditor, machine manager, and prompt engineer.
All evidence suggests the tornado of job postings focused on AI/ML will continue to rise rapidly across the U.S. federal government for years to come.
.svg)
.png)
Between 2008 and 2015, the various islands making up the U.S. territories received half the total media attention of equivalent-size states (4,936 vs. 10,138 articles).
An uptick in coverage of the territories in 2016 and 2017 was largely driven by the Puerto Rican debt crisis and the devastation of Hurricane María in the Caribbean.
North Korea’s 2017 announcement that Guam would be the target of its nuclear missile program also contributed to increased coverage that year.
In 2018 The New York Times reported heavily on the destruction caused by Typhoon Yutu in the Northern Marianas Islands.
The historic election of 2020 and the COVID-19 pandemic exacerbated the disparity in coverage between states and territories.
While native residents of the territories are generally U.S. citizens (except in American Samoa where they are U.S. nationals), they do not have the right to vote in general elections. Coverage about COVID-19 levels in different states and counties often excluded the territories.
The Enormous Federal Data Disparity
Federal data collection largely stops short of U.S. territories. Over the next 10 years, the Census Bureau will release approximately 264 key datasets for the 50 states. But for the territories of Guam, the Northern Marianas, American Samoa and the Virgin Islands, the Census Bureau will release only three total datasets over that same period: one decennial count and two economic surveys.
All told, the data collected by the Census Bureau will help direct at least 2.8 trillion dollars annually to 353 federal-assistance programs. While data on the 50 states helps the government direct funding where it’s needed most, a lack of territory data forces officials to operate in the dark.
For its most populous territory, Puerto Rico, federal data collection is a little better. The Census Bureau conducts an annual “Puerto Rican Community Survey” for the region’s 3.2 million residents. But the resulting estimates don’t use the same rigorous control methods as the “American Community Survey.” And data is only available on the county level instead of more specific geographies like zip codes and census tracts.
A Blindfold for Local Officials
Around the turn of the 20th century, the U.S. expanded its colonial influence over seas. Long left to the rule of the U.S. Navy, the nation’s territories were neglected by the government that claimed to rule them. This neglect hampered the development and assessment of the regions’ social programs. In recent years, a lack of federal data has hindered the ability of island territories to respond to disasters like the COVID-19 pandemic and extreme weather.
For example, in 2018, Typhoon Yutu devastated the Northern Mariana’s islands. By 2020, the recovery effort had just gotten underway when the spread of COVID-19 crushed the region’s critical tourism industry. When the federal government asked the region’s department of labor to estimate how many workers lost their jobs during the pandemic, they had no idea.
Speaking to the Honolulu Civil Beat, the head of Northern Mariana’s labor department Vicky Benavente said, “This is one lesson we learned. Data is so critical for justifying our asks to the federal government.”
State governments had ready access to reliable data. They used monthly reports from the Current Population Survey to monitor pandemic-induced rises in unemployment. Working without this data, the Northern Marianas government had to rely on a survey of employers conducted every two years. By 2021, so many businesses had shut their doors that few were left to reply.
About the Data
USAJobs is the official job portal of the U.S. federal government. It provides a comprehensive listing of job vacancies available across agencies, job fields, and locations. The Citizen Codex team extracted a historic data pull of all open job postings, from 2018 onward, and parsed the data using a simple keyword search across job descriptions. Keywords included: “data science,” “artificial intelligence,” "advanced analytics," and “machine learning.”
.png)
.svg)
.svg)
.svg)
.svg)