Technology & innovation
The rush to build AI infrastructure is generating a wave of construction work and skilled trades jobs, but the long-term employment picture is more limited than the headlines suggest. Meta's new America's Workforce Academy, a $115 million first-year program, is the most visible example of Big Tech's effort to build the workforce pipeline it needs.
Meta's program offers paid training, travel, lodging, and a job offer to graduates in skilled trades tied to data center construction, with 2026 pilot locations in Louisiana, Ohio, Indiana, and Texas; a prior Meta fiber training program drew 35,000 applications in its first seven days.
U.S. data centers are expected to generate 4.7 million temporary construction jobs and roughly 697,000 permanent jobs to operate and manage facilities, according to a 2025 report from the American Edge Project, a policy advocacy group founded by Meta.
Permanent staffing at the most automated data center campuses runs about 25 to 40 operators per 100 megawatts; Meta's $10 billion Lebanon, Indiana campus will employ roughly 300 people once operational.
Amazon Web Services plans to invest $35 billion in Virginia data centers by 2040, creating at least 1,000 total new jobs across the state; an Ark Data Centers campus expansion in Ohio worth $136 million will create exactly 10 permanent jobs.
A Brookings Institution analysis of roughly 770 U.S. data center facilities found counties receiving their first large data center see total private employment rise 4% to 5% over five to six years, but gains depend heavily on having multiple facilities in the same area.
Almost half of state data center subsidy programs do not require job creation, and the average cost of data center "megadeals" is $1.95 million per job, according to Good Jobs First.
Read more via Fox News, CBS News, Quartz
Oracle reduced its global headcount by approximately 21,000 employees, or about 13%, in the fiscal year ending May 2026, paying $1.84 billion in severance, up from $374 million the prior year. The company cited management changes, product portfolio shifts, performance issues, strategic changes, and expanded AI use as reasons for the restructuring.
Oracle employed roughly 141,000 full-time staff as of May 31, 2026, down from about 162,000 a year earlier.
Cloud revenues for the full fiscal year reached $34 billion, up 39%; total revenue hit $67.4 billion, a 17% increase.
Cloud infrastructure revenue grew 93% in the fourth quarter alone.
Average employee tenure at Oracle is eight years, with just under a third having worked there at least a decade.
Read more via Yahoo Finance
The Trump administration is urging Meta to allow voluntary government review of its AI models, making it the last major U.S. AI developer without such an agreement. OpenAI, Anthropic, Google, xAI, and Microsoft have all submitted models to the Center for AI Standards and Innovation, the Commerce Department's AI safety group.
Meta said it hopes to sign the agreement soon, calling it part of advancing U.S. leadership on "robust and secure frontier AI."
On June 2, President Trump signed an executive order giving the government responsibility for AI reviews and asking companies to allow up to 30 days for evaluation before public release.
Anthropic's newest model, Fable 5, was pulled by the White House with less than 90 minutes' notice after a researcher identified a potential cybersecurity vulnerability; Trump subsequently said he no longer viewed Anthropic as a security concern.
It remains unclear who will lead the review process or what standards models will be held to.
Read more via The New York Times
A new EY study finds wide gaps in AI readiness among finance leaders, with just 21% rating their preparedness as leading or advanced. Only 5% said AI is fully integrated into their operations and actively driving business decisions.
Seventy-one percent of CFOs say traditional financial metrics are not sufficient to evaluate initiatives that combine people and technology.
Nearly half (47%) said their teams lack the ability to effectively measure value created by emerging technologies.
CFOs at companies with more than $10 billion in revenue were significantly more AI-ready: 25% described their capabilities as advanced and 10% as leading, compared with 12% and 2% at smaller firms.
Most CFOs are applying AI defensively, to fraud detection and risk assessment, with fewer than half using it for growth-oriented functions like forecasting or pricing strategy.
Read more via CFO Dive
Manufacturers across the U.S. are moving from AI experimentation toward agentic systems capable of autonomous planning, scheduling, and decision-making, but most don't yet have the data foundation to support them. According to Deloitte's 2026 State of AI in the Enterprise report, nearly three in four companies plan to deploy agentic AI within two years, but only one in five reported having the infrastructure to support it.
Eighty-seven percent of manufacturers had already launched at least one generative AI pilot, according to Deloitte's 2025 Future of Manufacturing report.
Gartner projected last year that more than 40% of agentic AI projects could be abandoned by 2027 due to unclear business value, rising costs, and implementation challenges.
Predictive maintenance and vision-based quality inspection are currently the strongest use cases, with clear ROI and relatively established sensor infrastructure.
The gap between operational and information technology remains a major constraint; clean, consistent, integrated data is a prerequisite for agentic AI that most plants don't yet have.
Experts describe the current state as "semiautomatic human-in-the-loop," with human validation still required for unplanned scenarios even where capable agentic systems exist.
Read more via Manufacturing Dive
A growing number of HR leaders are moving away from annual performance reviews toward continuous feedback models, and AI tools are making that shift more practical for time-pressed managers. Two Canadian HR executives described how AI note-taking and real-time documentation tools have changed their approach.
Among organizations already using AI in HR, 89% reported greater operational efficiency, according to a Society for Human Resource Management white paper.
Managers using AI note-taking tools can stay present in conversations rather than breaking to write things down; notes are shared with employees afterward as a transparency measure.
HR leaders caution that consent matters: employees should be informed when AI is recording a meeting, and personal identifying details should not be entered into AI tools.
Both executives emphasized that technology should enable more frequent human contact, not replace it.
Read more via HCA Magazine
As organizations accelerate AI adoption, the employees leading that charge, often top performers, are being asked to do more than anyone else, and HR leaders are worried about losing them. Eighty-eight percent of people leaders say retaining top talent is their biggest priority right now, according to a survey by wellness platform Wellhub.
Eighty-five percent of people leaders said they are using wellness programs to retain top talent.
AI is either amplifying burnout or amplifying performance, according to Wellhub's chief revenue officer, and organizations are trying to ensure they land in the second category.
Top performers are disproportionately at risk because they are often tasked with leading AI adoption, supporting colleagues in building new skills, and taking on additional oversight responsibilities simultaneously.
Read more via HR Brew
Internal dissent at Meta has spilled into public view, with employees describing a toxic and demoralizing work environment driven by CEO Mark Zuckerberg's all-in push on AI. More than 25,000 employees have been laid off since 2022, with another 8,000 cut this month, as the company redirects resources toward AI development and recruitment.
Workers in Meta's elite Applied AI unit told Wired the work is "soul-crushing," consisting largely of devising puzzles and coding problems to train AI models.
Zuckerberg announced Meta would monitor staff work to train task-automating applications, which employees understand will eventually replace them.
During a live-streamed company presentation, an unidentified employee delivered an obscenity-filled outburst about being "the company's bitch," apparently to widespread employee approval.
Meta CTO Andrew Bosworth acknowledged morale was near its worst in the company's 20-year history; Chief Product Officer Chris Cox called the environment "brutal" and compared it to "running a marathon in the middle of a hailstorm."
Zuckerberg's attempt to boost morale with a July hackathon invitation drew internal responses from employees saying they had no time or incentive to participate.
Read more via Inc.
A new NPR/Ipsos poll of 545 K-12 teachers found that nearly three in four believe AI has larger implications for education than past innovations like the internet or computers. Most are using it themselves to save time, but a majority worry it is eroding students' ability to think independently.
Six in ten teachers said they have used AI to help with work tasks; a majority said it saves time, but 63% said that savings amounts to two hours or less per week.
A little more than half of teachers said AI is not being used in the classroom by students at all; about two in five said students use it at least once a week.
Fifty-four percent said AI makes it harder for students to learn critical thinking skills, and 55% said they believe AI is mostly just a shortcut for students to avoid doing work.
Nearly six in ten said AI is eroding trust between students and teachers; about four in ten have responded by requiring more assignments to be done by hand or completed in class.
Nearly eight in ten think schools should teach responsible AI use, but only about four in ten said their school offers professional development or training related to AI.
About half of all polled teachers said their school has offered no guidance on AI, or they are not sure what the guidance is.
Read more via NPR
Rivian's CEO quietly built a humanoid robotics company. Mind Robotics has raised more than $1 billion and expects to reveal its first product within a year. RJ Scaringe is keeping it separate from Rivian, a deliberate contrast to Elon Musk's approach of integrating robotics directly into Tesla. Scaringe expects robots to handle the simplest manufacturing tasks while humans handle work requiring higher-level reasoning and dexterity, and says manufacturers are currently dealing with an "extreme lack of labor." (CNBC)
Medicare's AI prior authorization pilot is causing delays and errors. The program, called WISeR, launched in January across six states and covers 13 services flagged as fraud-prone, including epidural injections. CMS says 88% of applications get an immediate approval; in practice, doctors report six- to eight-week payment delays and denials they attribute to AI hallucinations. One Phoenix pain doctor said as of May he hadn't been paid for nine epidurals. Vendors say humans make the final calls, but doctors aren't convinced. (CBS News)
Agentic AI loops may be the next big shift in how AI does work. Claude Code creator Boris Cherny made the case at Meta's @Scale conference last week: rather than managing individual AI agents on discrete tasks, developers authorize swarms of agents to run continuously in the background, endlessly improving code and filing pull requests. Cherny called the shift from agents to loops as significant as the shift from writing code by hand to using agents at all — though as TechCrunch notes, it is also a potentially expensive way to work, with no ceiling on token spend. (TechCrunch)
Anthropic is paying people to bring AI to nonprofits. Claude Corps, a new year-long fellowship, places early-career workers inside nonprofit organizations at $85,000 a year to help them integrate AI, with an initial $150 million committed to place 1,000 fellows at more than 400 organizations nationwide. No technical experience required — just be over 18 with under two years of full-time work experience. The first cohort of 100 starts in October. Notably, Amodei announced the program the same day he published an essay arguing AI-driven job displacement may be unavoidable and calling for universal basic income funded by taxes on AI companies. (Tom's Guide)
