Finding Work With AI
AI for DC
Hi
Welcome (back) to The Prompt. AI isn’t well understood, but we learn a lot in our work that can help. In this newsletter, we share some of these learnings with you. If you find them helpful, make sure you’re signed up for the next issue.
[Data] Finding work with AI
Tens of millions of adults in the US are turning to ChatGPT for help with a job search, mainly application prep and career exploration, according to early results of an analysis by OpenAI’s Economic Research team.
The team created a sample of more than 600,000 ChatGPT conversations by 13,000 de-identified adult users that happened between May 2024 and July 2025 and ran them through an automated classifier. An estimated 3.5% of the conversations were about finding work or moving through a job application process. Of the 13,000 users, about one third of them used ChatGPT at least once to help with a job search. Extrapolating these estimates suggests that about 40 million US adults used the tool at least once for help with a job search.
The analysis highlights the different ways people leverage AI to find a job, with résumé/cover letter preparation making up an estimated one-third of all job search-related conversations, career exploration/research nearly 30%, and help with communication with recruiters making up 20%.
Note: Percentages are probabilistic estimates based on automated classifiers; 10% of conversations were classified as requesting multiple capabilities.
The analysis also finds that older job-seekers use ChatGPT differently from younger ones. Older job-seekers use ChatGPT to leverage their career capital – being much more likely to use ChatGPT to rewrite their résumés and cover letters and to communicate with recruiters. Younger job-seekers use ChatGPT to build their career capital – they are much more likely to use ChatGPT to prepare for interviews, research open positions, and strategize career paths.
(See below for how American service members are using ChatGPT to help transition to the civilian workforce.)
Taken together, these findings suggest that people are using ChatGPT as a kind of low- or no-cost, always-on career center. It handles practical tasks like formatting résumés or drafting emails. It gives people a place to ask basic questions that might feel awkward with an interviewer. Overall, it helps break down a daunting process into discrete, manageable steps, making it less onerous to adapt to a changing job market. We’ll be incorporating these learnings into our forthcoming OpenAI for Jobs platform in 2026.
[Insight] Labor market gut-check
So far, AI’s early impacts on white-collar work are emerging unevenly. Explicitly AI-attributed layoffs amounted to less than 1% of all official layoffs and separations in Q1-Q2 2025. Some early-career jobs are clearly under pressure, posing a real challenge for younger workers. At the same time, AI is beginning to create new classes of work that draw on operations and judgment skills rather than advanced technical abilities, making them accessible to a broad range of workers.
On explicitly AI-attributed layoffs: Multiple studies find no broad labor-market downturn as of yet: Goldman Sachs reports “little empirical evidence” that AI is dragging down overall job growth, and the Yale Budget Lab reports no measurable nationwide displacement to date. The effect we can see right now is task-shift, not mass unemployment—AI is taking on drafting, summarization, coding, and documentation, while people move into review, troubleshooting, and complex decision-making.
On early-career jobs: Entry-level postings in software, marketing, and administrative support have contracted, and young workers in the most AI-exposed roles have experienced a 13% relative employment decline since generative AI’s adoption surge. We don’t yet have official numbers for AI-attributed layoffs for Q3-Q4 and will update this assessment as data comes across.
On new, AI-driven classes of work: We’re seeing rapid growth in model-oversight, review, and orchestration roles. Walmart is consolidating dozens of agents into “super-agents,” requiring human routing and governance. McKinsey now operates 12,000 AI agents alongside 40,000 consultants. Financial firms like Goldman Sachs, Morgan Stanley, and BlackRock are deploying AI systems that depend on human auditors, workflow designers, and evaluators—new roles that draw on operations and judgment skills rather than advanced technical abilities, making them accessible to a broad range of workers. (Check out the OpenAI Global Affairs report on the new kinds of roles we see in the AI era.)
This aligns closely with a new Washington Post analysis showing AI displacing some white-collar roles while simultaneously fueling demand for data labelers, prompt engineers, and model-oversight staff—an “AI partner economy” reminiscent of previous tech-shifts in which new, higher-skill ecosystems grow around the technology.
Taken together, our research alongside external research points to a labor market being reconfigured, not hollowed out. Some traditional entry points into white-collar careers are shifting, but entirely new pathways—AI operations, agent-management, evaluation, workflow design, human-in-the-loop quality control—are scaling quickly across tech, finance, consulting, and retail. Whether this shift becomes a net positive will depend on how quickly workers, employers, and institutions adapt to jobs that increasingly revolve around directing, verifying, and amplifying AI. – OpenAI Economic Research
[News] AI for veterans
Hundreds of thousands of US service members transition from military to civilian life each year. To help with their transition, OpenAI this week announced we’re giving a free year of ChatGPT Plus to service members within 12 months of retirement or separation. The idea came from veterans who now work at OpenAI and used ChatGPT to help with their own transitions:
Daniel An, Security Engineering, OpenAI
Hi, my name is Daniel An. I am a Security Engineer at OpenAI and a former active duty Army Officer. I transitioned out of the Army in the summer of 2024. The process was complex and at times overwhelming, but ChatGPT played a critical role in making that journey manageable.
Anything related to transitioning out of the military can be a challenging subject, especially when you’re not ready to tell your peers or chain of command. Early on, I wasn’t even sure whether I wanted to leave. I used ChatGPT to ask questions I couldn’t ask elsewhere, freely discussing the pros and cons of staying in, weighing career options, finances, benefits, locations, and long-term goals for my family. This helped me make a clear and confident decision about my future, especially when I needed to explain to my family and chain of command why I am making this decision.
Once I decided to separate, I turned to ChatGPT to understand what the ETS timeline and paperwork would look like. Eighteen months out, it provided me with a structured overview of key milestones, from required briefings on the TAP checklists, UQR approval process, SkillBridge and out-processing. That gave me the clarity to ask the right questions to my Chain of Command and HR, and cross-reference and doublecheck everything with the official Army regulations. I was able to stay ahead of the game instead of reacting to them at the last minute or waiting for someone else to give me answers.
David Sperry, OpenAI for Government
I’m David Sperry, and I’m an AI Adoption Manager at OpenAI.
I enlisted in the Army right out of high school and served for 20 years in a non-technical role, moving from post to post around the world. When I left the military, I knew I wanted to work in technology but didn’t have the traditional background. My first opportunity came in customer success at a major tech company, helping a large public-sector customer migrate from legacy data centers to the cloud.
That role was exciting—and humbling. I brought two decades of leadership and problem-solving experience, but I was missing a lot of the enterprise IT context: acronyms, architectures, and motivations that everyone else seemed to know. I was learning cloud computing while also translating the culture of the military into that of big tech.
Six months in, ChatGPT was released. It instantly became my new best friend. I used it to fill every gap—asking it to explain technical concepts, get insight into customer situations, and translate technical jargon into plain English. It let me learn privately, quickly, and on demand. Over time, I became a power user. Hundreds of prompts turned into thousands, and I began to understand not just how to use AI, but how to think with it.
I started sharing what I learned—first with coworkers, then with customers. They wanted to know how to begin their own AI projects while staying within public-sector constraints. Few people then understood both the technology and the unique requirements and constraints of the public sector. I decided to become that bridge. I used AI to teach myself AI so I could help others do the same. It became a virtuous cycle: the faster and more I prompted, the faster I learned, and the more I could give back.
[Event] Equipping small businesses to be mighty
With the right AI tools, small businesses can punch above their weight, compete in new ways, and keep Main Streets vibrant. Three quarters of America’s smallest businesses (with 1-9 employees) see AI as important for their future, and of all sizes of business, they’re most likely to cite a lack of AI literacy as the blocker to adoption. OpenAI’s upcoming Small Business Jam – in partnership with DoorDash and SCORE – will help them use ChatGPT to work smarter and serve customers better.
On November 20, 1,000 small business owners in five cities will get a full day of hands-on instruction from OpenAI mentors on how to use AI to streamline everyday tasks, from marketing and scheduling to customer communication. Each business will leave with at least one AI tool they can put to use immediately to help their small teams do more with less.
The Jam is part of our broader effort to ensure AI’s economic upside reaches every corner of the country. And for those who can’t attend our upcoming in-person event, we will be hosting a virtual Small Business Jam on December 4. You can register here.
[About] OpenAI Forum
Explore Forum programming by and for our community of 60,000 AI experts and enthusiasts from across tech, science, medicine, education, government, and other fields.
3:30 PM – 4:35 PM EDT on Nov 13
[Disclosure]
Graphics created by Base Three using ChatGPT.










Hey, great read as always, it's truly fasinating to see how AI is helping so many people with job searching, kinda like how I use it to discover new authors and expand my reading lists, just blows my mind how versatile it can be.
The data here is genuinely exciting. Forty million people turning to ChatGPT for career support shows how quickly AI is becoming an equalizer, not a threat.
The age differences are fascinating, too. Older workers leveraging experience, younger workers building it. That’s exactly what a healthy transition looks like.
And the rise of “AI partner economy” roles feels like one of the biggest underreported shifts in the labor market. Oversight, orchestration, judgment; these are new paths, not dead ends.
Really appreciate this clear, grounded breakdown. Looking forward to seeing what OpenAI for Jobs unlocks next.