AI & Future of work - 25 May 2025
Will Veo 3 kill online courses? Will agents take over?
Welcome to FullStack HR, and an extra welcome to the 34 people who have signed up since last edition.
If you haven’t yet subscribed, join the 9300+ smart, curious and like-minded HR people by subscribing here:
Happy Sunday, everyone!
I know I’m late but as they say, better late then never.
Workday Faces Class Action over Allegedly Biased AI Hiring Tools
A California judge on 19 May let Mobley v. Workday proceed for every U S applicant aged 40+ whose résumé was filtered out by Workday’s screening algorithm, putting vendor-supplied AI squarely inside employment-law risk.
Why it matters for us in HR: Time to tightening the vendor contracts and scheduling quarterly bias audits? I think this might be the time to do so in order for us to show, not tell, that our models treat age as a protected factor, not a red flag. Ties back into last weeks article.
Read More →Tech Giants Cut 61 000 Roles as AI Reshapes Org Charts
By 23 May more than 130 tech companies, including Microsoft, Google, Amazon and CrowdStrike, had eliminated over 61 000 jobs while redirecting budgets toward generative-AI engineering.
Why it matters for us in HR: Layoffs justified as “AI efficiencies” are the new normal now I guess? Even if Microsoft was a bit vague on if it was due to AI efficiencies or just a “budget recalibration”.
Read More →Google DeepMind Ups the Bar with Gemini 2.5 “Deep Think”
At I/O on 21 May, Google rolled out (amngst other things) an updated and wild Veo 3, Gemini 2.5 Pro Deep Think and 2.5 Flash, adding native audio output, a million-token context window and stronger reasoning for tough math and code.
Why it matters for us in HR: Veo 3 has been the mindboggling release for me here. Sure, there are a lot of things to unpack but the ability to generate video AND sound, instantly. Think about how that will impact L&D.
Read More →Anthropic Launches Claude Opus 4 and Sonnet 4
More releases. Launched Thursday, the Claude 4 family sets new coding and agent benchmarks and ships with built-in tool use plus ASL-3 safety protocols.
Why it matters for us in HR: Initial tests of Claude 4 showcases one of the best models out there. But what also has sparked controversy is that when testing the model, the model tried blackmailing (!!!) an engineer that had the intention to shut down the model. Anthropic is leading the way in safety-testing and it’s marvellous that they do it in the open. More should hopefully follow to make us better buyers. (Once again, see point 1)
Read More →SAP Adds Real-Time, Graph-Grounded Answers to Joule
At SAP Sapphire on 20 May, SAP made its Joule copilot omnipresent across ERP modules and tapped Perplexity’s answer engine for chart-ready insights.
Why it matters for us in HR: This marks a shift from dashboards to conversations. HR is no longer just the owner of people data, we’re expected to contextualise and interpret it instantly. Leaders won’t wait for monthly reports when they can get real-time insights through AI. Our role is evolving toward being sensemakers in a world flooded with automated answers.
Read More →Microsoft Build Showcases Multi-Model Agent Platform
Yes this was THE week of launches because on 19 May Microsoft previewed Azure support for Anthropic, OpenAI and xAI models plus low-code tools that let multiple Copilot agents collaborate on long-running tasks.
Why it matters for us in HR: Copilot Studio lowers the barrier to building digital coworkers that handle onboarding, policy compliance and pulse surveys. HR tech teams should run rapid prototypes to quantify time saved, then partner with risk and change-management to scale responsibly.
Read More →PwC Survey Shows 88 Percent of Executives Raising AI Budgets
PwC’s May poll of 300 U S C-suite leaders found 88 % plan to boost AI spend within 12 months and 75 % expect agents to transform work more than the internet did.
Why it matters for us in HR: Investment is surging faster than operating models can catch up. That disconnect is an opportunity for HR to lead, not follow, in organisational redesign. This isn’t just an IT trend. It’s a culture, capability, and leadership challenge that we in HR are uniquely positioned to shape.
Read More →HBR: “Agentic AI Is Already Changing the Workforce”
A 22 May analysis argues that digital-labor markets of autonomous agents could reach trillion-dollar scale and redefine what counts as a “qualified” worker.
Why it matters for us in HR: As said numerous times here before: We’re entering an era where “workforce” no longer just means people. Agents are becoming part of the organisational structure. That means new job architectures, new performance metrics, and eventually, new kinds of accountability. HR must begin planning for hybrid teams, of humans and machines, not as a future concept, but as an immediate reality.
Read More →Ethan Mollick on Substack: “Making AI Work: Leadership, Lab, and Crowd”
In his 22 May post, Mollick offers a three-part blueprint—executive vision, safe sandbox, employee experiments—for scaling AI across organisations.
Why it matters for us in HR: This is part of what was delivered at Sana’s AI event a couple of weeks back. Mollick’s framing hits the sweet spot between control and creativity. AI adoption thrives when employees are trusted to experiment, within clear boundaries. HR should be the architects of those boundaries, not the blockers. This is a moment to embrace grassroots innovation and make room for play inside work
Read More →HR Dive: Almost Seven in Ten Companies Now Use AI at Work
A 20 May OwlLabs survey reported that 67 % of businesses have AI in daily workflows, and employers are actively encouraging use for the first time.
Why it matters for us in HR: This data shows how fast the shift is happening, and how uneven it is. Enthusiasm is outpacing education. Without structured support, employees will develop their own shadow practices. HR must lead the charge in building confident, competent AI fluency across the organisation, before the gaps become risks.
Read More →