Xi Song, Jennie E. Brand, Sukie Xiuqi Yang and Michael Lachanski examine how AI can speed the decline of some occupations and make it harder for affected workers to move into better jobs (AEA Papers and Proceedings, May 2026).
Why I recommend it: A short conference paper, so the evidence is brief. "Mobility trap" is the authors' framing, not settled fact.
#ai and jobs#economic mobility#economics#occupational change#research
FinanceBuzz article by Josh Koebert, syndicated on MSN, listing nine fields it argues are shrinking: journalism, computer programming, data entry, telemarketing, photography, translation, bookkeeping, legal assistance and customer service. It cites Bureau of Labor Statistics projections — 4% fewer journalism jobs, 6% fewer programming jobs and 6% fewer bookkeeping jobs by 2034 — and points at information security analysts as a growing alternative with a 2024 median salary above $124,000.
Why I recommend it: Useful as a prompt to check your own field's BLS projection — not as career advice. Three things I would not take from it: the headline conclusion is the writer's opinion, not the BLS's; a projected national decline says nothing about whether good jobs exist in your city; and the page is packed with affiliate links to insurance and money products, which is how it earns. Look up the BLS Occupational Outlook Handbook entry for your own job title and read that instead.
#ai and jobs#automation#bureau of labor statistics#career change#career planning#cybersecurity careers#declining industries#job markets#job outlook#job search#research#salary
Manoush Zomorodi (NPR's TED Radio Hour) coins p(joy) as a deliberate counterpoint to p(doom), the probability-of-catastrophe number traded around AI circles: instead of estimating disaster, estimate whether satisfying moments are increasing in your life and what reliably creates them. The second half is a long interview with Brian Elliott of Work Forward on knowledge work right now — identity loss as machines take a skill you spent decades building, cognitive overload from supervising AI agents, and executives privately saying their best people and heaviest AI users are the likeliest to quit. Elliott's summary: human-led, AI-enabled, because nobody's competitive advantage is being average faster.
Why I recommend it: Read this the next time an AI doom number is quoted at you as though it were measured. The piece cites Gizmodo's plain point that no one at the frontier labs has an actual method for producing a p(doom) figure. The useful part is Elliott's three patterns in companies getting value from AI: a scorecard covering business impact, quality and employee engagement rather than tool usage; aiming at teams instead of individuals; and leaders using the tools in public, mistakes included. Free to read in full, no paywall.
#ai and jobs#burnout#career planning#future of work#job markets#knowledge work#leadership#newsletter#p(doom)#p(joy)#workplace
US Census Bureau working paper (CES 26-56, September 2026) by Cody Orr, Lee C. Tucker and Lawrence Warren, using administrative records covering about 29% of US bachelor's degrees conferred 2016-2024. Graduates in the most AI-exposed tenth of majors saw their chance of being employed in the quarter after graduation fall five percentage points, and first full-quarter earnings fall thirteen percent, starting immediately after ChatGPT's release in late 2022. The earnings hit is comparable to graduating into a large recession. About half came from lower pay inside the same industries, the rest from graduates shifting into lower-paying sectors such as restaurants and retail. The effect shrinks to about five percent two years out but does not disappear for the most exposed majors.
Why I recommend it: This is the strongest evidence yet that AI has already moved entry-level pay, because it uses actual wage records rather than employer statements or surveys. Two things to hold onto: the paper says plainly it has not been through Census Bureau review and is not an official position, and exposure is measured by what a major typically leads to, not by whether any particular employer used AI. So it tells you which fields got harder to enter — not that a machine took a named job.
#ai and jobs#automation#career planning#census bureau#college graduates#earnings#economics#entry level#job markets#labor market#research
Pew Research Center short read, 18 August 2026, on a survey of US adults conducted 22-28 June 2026. 52% now say they are more concerned than excited about AI in daily life, up from 37% in 2021. Among adults aged 18-29, 55% are more concerned than excited and 11% more excited than concerned, and 73% think AI will mean fewer US jobs over the next 20 years, up from 61% in 2024.
Why I recommend it: Useful when a client says the worry is just in their head — it is not, and the numbers are free to download as a spreadsheet. Read it as what people expect, not as what has happened to employment: it measures opinion, not job counts.
Stanford Digital Economy Lab working paper by Bharat Chandar and Bouke Klein Teeselink, separating what happens to jobs after a firm adopts generative AI into a company-wide productivity effect and changes in what specific kinds of workers are asked to do.
Why I recommend it: Free, with the full PDF on the page. It is a working paper dated 21 September 2026, meaning it has not been through peer review yet, so read it as early evidence. Useful because it does not answer 'will AI take jobs' — it asks which workers get asked to do more and which get asked to do less.
Chapter 3 of Pew's April 2025 report comparing US adults with AI experts on what AI will do over the next two decades. 56% of the experts surveyed expect AI's impact on the US to be positive, against 17% of the public; 35% of adults expect a negative impact, against 15% of experts. The gap is widest on work and money: 73% of experts think AI will positively affect how people do their jobs versus 23% of the public, and 69% versus 21% on the economy, with a 40-point gap on medical care (84% versus 44%). Experts and the public broadly agree on the risks to democracy and journalism: only 11% of experts and 9% of the public expect AI to help elections, while 61% of experts and 50% of the public expect harm. Gender splits are large among experts — 63% of male experts predict a positive impact versus 36% of female experts. Free to read, with methodology and appendix tables.
Why I recommend it: Use this when someone tells you 'the experts say AI will be fine at work' — the numbers show experts and the public are describing two different futures, and the widest gap of all is about jobs. Read it with two limits in mind: the fieldwork was in 2024 and published 3 April 2025, so it predates a lot of what has happened since, and Pew's 'AI experts' are people who published or presented at AI conferences, many of them employed by companies building AI, which is exactly the group the optimism gap belongs to.
Interdisciplinary Stanford research institute studying how digital technology and AI change work, productivity and shared prosperity, with public papers and data.
From the site: The Stanford Digital Economy Lab is an interdisciplinary research institute shaping a future where technology drives human well-being and shared prosperity.
Why I recommend it: Free to read. Useful when you want measured research on AI and jobs instead of headline claims — check the publication date on each paper, the field is moving fast.
IEEE Spectrum article on how AI is shifting entry-level engineering work toward higher-order thinking, review and collaboration skills, and what recent graduates can do about it.
From the site: How can recent grads navigate a job market transformed by AI? Learn how to make AI work for you, not against you.
Why I recommend it: Free to read on IEEE Spectrum. Practical if you are early-career: it describes what junior work is turning into rather than predicting job counts.
