Technology & innovation
A new Payscale report finds that employers are redesigning jobs around AI faster than compensation systems can keep up, creating inconsistent pay decisions and real retention risk. The full report drops in November; this is a preview.
Nearly half of employers (49%) say their salary structures haven't kept pace with AI, and 48% say current market benchmarking no longer reflects what roles actually require.
AI job postings have grown 8.7 times in less than five years, according to Lightcast data, but 41% of employers say they can't find candidates with the AI skills they need.
Pay strategies are all over the map: 58% of employers pay or plan to pay a premium for AI skills, while 23% now treat AI fluency as an unrewarded baseline and 13% haven't decided on an approach at all.
More than half of employees (56%) believe they should be paid more for developing AI skills, creating retention risk for employers who upskill workers but don't compensate them accordingly.
Three-quarters of employers (74%) plan to invest in AI-related upskilling over the next 12 months.
AI skill demand is outpacing employers' ability to price it. The result is inconsistent pay decisions at a moment when employees are investing in AI skills and expecting a return."
Read more via Payscale
A PayrollOrg survey of more than 9,000 workers finds deep employee skepticism about AI's role in pay processing, even as payroll technology is still in early adoption stages.
Forty-six percent of respondents said they are uncomfortable with AI being used in payroll processes; only 25% said they were somewhat or very comfortable.
The findings echo PayrollOrg's 2024 survey, when 39% were uncomfortable with AI calculating their pay and 53% were uncomfortable with AI answering payroll questions instead of a person.
Just 11% of payroll teams are currently using AI features in their technology, and only 25% of organizations have moved meaningful AI experiments into production.
Read more via PayrollOrg
A new Gartner report finds that finance leaders are concentrating AI investment on quick-return productivity tools while underinvesting in more complex use cases that take longer to mature but deliver greater business value.
Simpler tasks like data extraction and accounts payable automation deliver returns within nine to ten months; more complex use cases like forecasting and insight generation take longer but can improve decision-making and support revenue growth.
Low AI literacy has emerged as the top barrier to finance AI success, overtaking talent acquisition, as agentic coding tools have made technical barriers lower.
Gartner recommends giving finance employees hands-on AI experience through project assignments, sandbox experimentation, and short on-the-job activities rather than formal training alone.
Read more via Gartner
As AI costs mount, corporate tech leaders are experimenting with employee token limits, but finding that setting the right limits is harder than it sounds. Tokens are the quantifiable chunks of information AI models process, and costs vary by model and provider.
TIAA started with open access for all employees, then moved to function-by-function token limits, with a review process for employees who request increases; many requests are denied, with employees redirected to optimized prompt libraries instead.
Carvana opted against hard limits, instead setting monthly "expectations" by dollar value with weekly tracking and a review triggered when employees approach the threshold.
Principal Financial Group uses hard limits but made the increase request process quick and easy; the company initially set limits too low, causing some employees to ration their AI usage.
The core tension: limits that are too tight deter adoption, while limits that are too loose allow waste.
What I don't want people to do is stop using AI because they feel like they're going to get in trouble."
Read more via The Wall Street Journal
Check out the recent Need to Know SPOTLIGHT on AI Tokens
A Korn Ferry survey of more than 16,000 professionals across 11 countries finds that the proliferation of AI tools is adding to employee workloads rather than reducing them.
Sixty-two percent of respondents said their workloads had already increased significantly in recent years, before accounting for AI; 52% said AI tools have increased the number of tasks expected of them.
The dynamic reflects a well-known "J-curve": an initial productivity dip before employees get comfortable enough with new tools to realize gains.
Generic training and upskilling programs often compound the problem by adding more demands on already stretched employees; Korn Ferry recommends carving out dedicated learning time and making training role-specific.
Fifty percent of frontline workers worry AI will replace their jobs, but among those whose employers already use AI, 64% report a positive impact.
Maryland Gov. Wes Moore released an AI framework this week aimed at protecting workers as the technology reshapes jobs, as states continue to fill a regulatory vacuum left by federal inaction.
The framework calls for bringing together stakeholders to support workers through skills transitions, establishing a labor-management committee to work with unions on AI integration, and pursuing regulation of the "most concerning" aspects of AI.
Maryland already has one law on the books regulating employer use of facial recognition software during job interviews.
Six states and New York City have already passed laws specific to AI in hiring; California recently passed bills prohibiting AI surveillance of employees' emotional states and barring employers from relying solely on AI when firing or disciplining workers, both awaiting Gov. Newsom's signature by September 30.
Read more via HR Dive
A new analysis from outplacement firm Challenger, Gray & Christmas finds that the spread of AI notetaking and summary tools in hiring is generating a permanent record of candidate interactions that many employers aren't equipped to manage.
Popular web conferencing platforms including Zoom, Microsoft Teams, and Google Meet all offer transcription and AI summary features; recruiting platforms increasingly include tools that produce interview notes, fill out scorecards, and generate follow-up questions.
A hiring manager who reads an AI-generated summary without watching the underlying interview may be relying on a document that omits context, nuance, or contains outright errors.
A job seeker's mistake could be recorded permanently. An interviewer's insensitive — or even illegal — question could also be captured."
AI notetakers have raised compliance concerns around recording consent laws and workplace monitoring regulations.
Just over 90% of HR managers surveyed by Paylocity said they actively use AI somewhere in their recruitment process; four in five also said they were actively managing at least one issue with their AI tools.
Toyota is moving aggressively into humanoid robotics, with plans to invest $6.42 billion annually starting in 2028 to deploy 400,000 robots across its manufacturing operations, while insisting the push is about coexistence rather than replacement.
The company plans to place 150,000 robots in its own automotive plants and 250,000 at group company facilities; it has already begun deploying its ELEY humanoid robots, which roll on wheels and learn tasks from human workers wearing finger-based “jigs.”
Toyota's human workforce spans 18,000 veteran workers across 60 factories worldwide, including 11 U.S. manufacturing plants.
Hyundai plans to begin deploying Atlas humanoid robots developed by Boston Dynamics in its plants starting in 2028, with a goal of up to 25,000 robots across Hyundai and Kia facilities worldwide.
China deployed 2 million industrial robots in 2024; the U.S. ranked third with 34,200, according to the International Federation of Robotics.
Read more via Ars Technica, Nikkei Asia
Meta's Muse became the No. 1 app in Apple's App Store within two weeks of launch. The consumer AI agent can book appointments, make phone calls, negotiate purchases, and manage finances — but requires access to bank accounts, email, and calendars, and Amazon has already blocked it from shopping on its site. (The New York Times, The Wall Street Journal)
Anthropic's biology lab made a discovery in 21 hours that the company says has properties reminiscent of CRISPR, the gene-editing technology. Claude searched through data using roughly 950 agents and 210 million tokens to identify a previously unknown enzyme system in bacteriophage DNA; all physical lab work was performed by human scientists. Anthropic CEO Dario Amodei acknowledged the work builds on prior research, including from a Stanford team that found a similar system, and said the broader research community will need to validate the finding. (Nature)
A new Google feature that can call businesses on your behalf. Google is testing a feature called "Call for Me" that lets Gemini make phone calls to businesses on users' behalf, handling automated menus, waiting on hold, and conducting the conversation while users follow along via live transcript. The feature is rolling out initially to Pixel 11 owners with a Gemini subscription in the U.S. and builds on earlier experiments including "Ask for Me," "Hold for Me," and "Direct My Call." (TechCrunch)
