The federal government just said PR specialists, web developers and hotel workers now belong on its official list of jobs artificial intelligence could shrink. That's a meaningfully broader warning than the one it issued a year ago.
The agency that forecasts how many jobs the American economy will add over the next decade just said PR, web development and hotel work belong on its list of occupations artificial intelligence could shrink. Bloomberg reported the expansion on September 1, 2026, in its AI Today newsletter. The Bureau of Labor Statistics widened the AI exposure categories it publishes alongside its 2025 to 2035 Employment Projections. The new roster pulls in public relations specialists, web developers and hospitality roles. It already covered paralegals and legal assistants, graphic designers, interpreters and translators, procurement clerks, sales representatives, administrative assistants and customer service representatives. That's a much wider net.
Why these three jobs #
The BLS didn't invent this ranking out of thin air. Its exposure categories page explains that the agency combined several outside measures of AI exposure into four tiers, ranking every detailed occupation it tracks by how much of the work AI tools can plausibly touch. Outside models built by researchers including those at Anthropic and Microsoft have separately flagged public relations work as highly exposed. That's a real shift. A year ago, PR was treated as a craft insulated by relationships and judgment, not a job threatened by a script.
Web development lands on the list because so much of the job, template layouts, boilerplate code, routine front-end builds, now runs through AI coding assistants and site builders that didn't exist in a usable form three years ago. Hotel work is different. Front-desk check-in, booking changes and basic concierge questions are increasingly handled by kiosks and AI chat systems, not because hospitality itself got automated, but because the transactional slice of it did. The BLS isn't saying these professions disappear. It's saying the tasks inside them that used to require a person increasingly don't.
The layoffs already underway #
This lands on top of data Bloomberg already reported this year. In July, Bloomberg found that tech and finance sectors were shedding about 28,000 jobs a month, and tied the decline to accelerating AI adoption. The numbers were already grim. Challenger, Gray & Christmas has tracked roughly 102,000 AI-attributed layoff announcements in 2026, with tech accounting for close to a third of them. Census Bureau data cited in that same reporting showed 39.7% of information companies and 33.9% of finance and insurance firms already use AI tools, against a 19.8% national average. The BLS list expansion isn't a new trend. It's the government catching up to one Bloomberg had already been documenting.
Google DeepMind's Safety Team Warns Job Seekers Its Own AI Hiring Filters Misfire Bloomberg reports that Google DeepMind's AGI Safety and Alignment Team tells job applicants to submit a special workaround form because it doesn't trust Google's own AI hiring filters. The twist: Google markets that same AI screening technology to corporate clients as a selling point. - AI hiring bias in recruiting tools - why AI filters reject qualified candidates
For founders, that catch-up matters more than the list itself. An informal decision startups have been making quietly, skip the PR hire, run the launch through Claude and a comms freelancer, skip the junior web developer, build the marketing site in Framer, now has a government classification sitting behind it. That's not permission you needed. But it does change the conversation with a board member who asks why a headcount line item wasn't filled. Treat these BLS categories as a budgeting input, not a moral verdict. That's the test now. If a role sits inside a high-exposure bucket, the honest question isn't whether to hire it. It's whether the person you hire is doing judgment work, strategy, relationships, editorial calls, or production work a model already handles at a fraction of the cost. Not every AI-exposed sector is shrinking. The BLS still projects the U.S. economy adds 5.9 million jobs between 2025 and 2035. Professional, scientific and technical services, the bucket that includes AI consulting and development work, is projected to grow 8.6%, the third-fastest of any sector. Growth isn't the problem. Sales and related occupations are projected to shrink 1.4% over the same stretch, which the BLS ties directly to AI's incorporation into the sales process. The list got longer. So did the gap between roles built around judgment and roles built around repeatable tasks. That's where hiring decisions get made.
Also read: Baidu Says Its AI Business Is About to Earn as Much as Search Does • Wall Street Banks Are Pressing Law Firms to Cut Fees Because of AI • Abliteration.ai Now Sells an AI Model With No Safety Refusals for Hackers