Subscribe

The most interesting stories in HR & AI weekly

The Rise of “Robot Relations”

As companies deploy more AI agents and automated systems into everyday work, The Tech Buzz argues they may need dedicated capabilities to manage human-AI collaboration, transparency, conflict, and accountability. Why it matters: HR has an opportunity to play a central role in the lifecycle management and governance of AI agents, from defining roles and access to monitoring performance, managing escalation paths, and determining when agents should be changed or retired. Reality Check: The article’s “Robot Relations department” framing is more provocative than proven, but the underlying governance questions are real and could increasingly sit at the intersection of HR, IT, legal, and risk.

Microsoft Finds AI Tool Access Alone Doesn’t Drive Transformation

Drawing on hundreds of internal AI efforts, Microsoft says broad tool access by itself produced limited impact, while larger gains came when teams redesigned workflows around specific business outcomes. Why it matters: Microsoft’s experience reinforces that AI value may depend more on workflow and role redesign than on adoption rates alone. Reality Check: The results come from Microsoft’s own internal programs and selected use cases, so they should be treated as case examples rather than evidence of typical enterprise outcomes.

Researchers Used Claude to Break into OpenAI Systems

Security researchers reportedly used Anthropic’s Claude to help identify and exploit vulnerabilities in OpenAI systems, demonstrating how advanced AI models can accelerate offensive cybersecurity work. Why it matters: As AI lowers the skill and time required for sophisticated attacks, companies may need to rethink both cyber defenses and controls around how powerful models are used internally. Reality Check: This was a researcher-led security exercise, not evidence that Claude autonomously carried out a real-world attack, and the broader risk depends heavily on human direction and access. Related: Gemini hacked three companies in first known breakout by Google’s AI.

Trump Announces New AI Task Force and Plans to Appoint AI Czar

President Donald Trump said he will create an “AI Force” and appoint an AI czar to oversee the technology while emphasizing that the administration does not want to slow industry growth. Why it matters: The move signals a more centralized federal approach to AI policy as debates intensify over how much oversight the technology requires. Reality Check: The administration has not yet named the czar or provided details on the task force’s authority, structure or enforcement powers. Related: Obama Calls for Stronger Federal AI Regulation.

Newsom Pushes Independent AI Audits and a Frontier-Model “Kill Switch”

California Gov. Gavin Newsom issued an executive order accelerating independent oversight of frontier AI systems and directing experts to develop proposals for ongoing third-party audits and an emergency shutoff mechanism for advanced models. Why it matters: California is moving toward a more formal oversight model that could influence how AI companies document, test and govern high-risk systems. Reality Check: The order starts a policy-development process rather than immediately imposing a kill-switch requirement, and key details will depend on future recommendations and implementation.

Microsoft AI CEO Warns Against Treating AI Models Like People

Microsoft AI CEO Mustafa Suleyman argues that AI systems should not be treated as if they have consciousness, feelings, or rights, warning that “model welfare” could encourage anthropomorphism and complicate AI safety. Why it matters: As AI agents take on more work inside organizations, HR leaders will increasingly need to determine how these systems fit into workforce planning, accountability, performance, access, and lifecycle governance. Reality Check: Treating AI agents as part of the workforce for governance purposes doesn’t require treating them as people, the more immediate HR challenge is establishing clear ownership and controls for agents from deployment through reassignment and retirement.

CHRISTUS Health’s AI Front Door

CHRISTUS Health deployed NOAH, a Leena AI assistant, to help 51,000 employees navigate an HR environment with 149 applications, where answering a single question could previously require searching across three or four systems. Key insight: The case shows how AI can create a simpler employee experience without first replacing or consolidating a complex HR technology stack.

Culture Amp data from roughly 112,000 employees found that 85% are encouraged to experiment with AI, but 42% say leaders have not clearly explained how AI connects to company goals. Why it matters: The next AI adoption challenge may be less about getting employees to use the tools and more about helping them understand where AI fits into the business and their own work. Reality Check: The findings come from organizations already using Culture Amp’s AI Effectiveness survey and likely represent companies that are further along on AI adoption than the broader market.

A study of 890 supervisor-employee pairs found that workers who were more accepting of generative AI were more likely to engage deeply with AI-generated knowledge, which was associated with stronger supervisor-rated service performance. Why it matters: The findings suggest AI adoption may depend not just on access and training, but on whether employees actively interpret, integrate and build on what AI produces. Reality Check: The study is observational, so it shows associations rather than proving that greater AI acceptance causes better performance.

OpenAI research found that work-related ChatGPT use outside employees’ own occupations nearly doubled from 13.1% to 25.9% between April and July 2026, suggesting AI may be reshaping what people actually do before formal roles change. Why it matters: Workforce planning may increasingly need to track how tasks migrate across traditional job boundaries, not just headcount and job descriptions. Reality Check: The study is based on observational data from ChatGPT Business users and doesn’t establish whether broader roles translate into greater productivity.

Among workers concerned about skills erosion, three in four say AI has already weakened at least some of their skills, while 46% of CHROs also rank skills erosion among their top concerns. Why it matters: The findings reinforce why critical thinking and problem framing are becoming more important as AI takes on more routine cognitive work. Reality Check: These are self-reported perceptions of skills erosion, not objective measures of whether employees’ capabilities have actually declined.

A proposed class action alleges Workday’s AI hiring tools disproportionately screened out Black, female, over-40 and disabled applicants, raising a broader question of whether responsibility for discriminatory outcomes sits with the employer, the technology vendor or both. Why it matters: This may be the clearest US test yet of whether discrimination allegedly occurring upstream in an AI screening system can generate class-wide liability, even when employers retain formal authority over the ultimate hiring decision. Reality Check: The court has allowed the claims to proceed but has not found that Workday discriminated, and Workday denies wrongdoing, arguing its tools assess job qualifications rather than protected characteristics.

Employment attorneys warn that HR teams using public AI tools for hiring, terminations or workplace investigations may create chat histories that are discoverable in litigation and could jeopardize attorney-client privilege. Why it matters: HR leaders may need clearer guardrails around when employees can use AI for sensitive employment decisions and which tools are approved. Reality Check: Enterprise AI platforms can reduce confidentiality risks, but even closed-system chats may still be discoverable if the underlying communication was never privileged in the first place.

Brilliant AI, formed from Crosschq, ApplicantX and Traitify, launched a platform that combines interviewing, assessments, candidate verification and quality-of-hire analytics, claiming its models draw on more than 200 million hiring outcomes. Why it matters: The launch reflects a broader shift in recruiting tech from automating individual steps toward connecting hiring decisions with post-hire outcomes. Reality Check: The performance claims come from the vendor itself, and the release provides limited independent evidence that its “hiring intelligence” approach improves quality of hire across customers.

HiBob is integrating The Josh Bersin Company’s Galileo research and frameworks into Bob AI, allowing agents to combine external HR guidance with company-specific workforce data for areas including job architecture, compensation and performance management. Why it matters: The launch points to an emerging HR-tech model in which AI agents draw on both internal company data and external knowledge sources to support workforce decisions. Reality Check: The offering is still early, with full availability planned for 2027, and HiBob has not yet provided evidence that adding external HR content leads to better workforce decisions.

Thanks for reading. If you have comments or feedback just respond to this email.

Until next week,

Mike & Adam

HRBusinessPartner AI is published by PotomacHarbor. Follow us on LinkedIn.

Love HRBusinessPartner AI? Share your referral link below to get HRBusinessPartner AI swag!

{{ rp_refer_url }}

Someone forward you this email? Subscribe for weekly updates.

Want to advertise in HRBusinessPartner AI? Send us an email (reply to this message).