AI is changing LinkedIn on several fronts at once. It is taking on more of the recruiting workflow, powering the systems that decide what deserves distribution in the feed, and reshaping the wider Microsoft organization around the future of work.
This week brings a new generation of LinkedIn’s recruiting agent, a rare look inside the technology used to suppress AI slop, and a major leadership change after nearly two decades.
LinkedIn’s Hiring Assistant Is Becoming a Full Recruiting Agent
LinkedIn has announced Hiring Assistant 2, the next generation of its AI agent for recruiters, with stronger reasoning, memory, personalization, and the ability to manage a broader part of the hiring process.
The timing reflects a difficult hiring environment. LinkedIn says global hiring is down 30%, while applications per applicant are up 30% compared with pre-pandemic levels. At the same time, 64% of U.S. talent acquisition professionals say it is becoming harder to know what is real and who they can trust when assessing candidates.
Hiring Assistant 2 is designed to help recruiters cut through that volume. It combines candidate verification, ATS information, screening responses, work experience, and other signals in one place. It can also learn how individual recruiters and teams hire, then use that context to recommend stronger candidates and manage the wider talent pool.
The bigger shift is that LinkedIn is moving beyond AI as a sourcing tool. Hiring Assistant is becoming a more agentic system that can manage entire workflows, while recruiters are pushed toward the parts of hiring that still require human judgment, such as understanding business needs, evaluating capabilities, and building relationships with candidates.
That fits LinkedIn’s broader view of recruiting. As skills change faster and AI makes applying easier, the value of talent acquisition increasingly comes from strategy and trusted signals, not simply processing more applications.
LinkedIn Says Its AI Slop Detection Now Reaches 94% Precision
LinkedIn has shared new details about the technology behind its fight against low-value, repetitive AI content.
The platform says its classifiers now cover all posts distributed beyond a member’s immediate network, with 94% precision when identifying content LinkedIn classifies as AI slop. By August, members were seeing 40% fewer views on this type of content compared with only a few weeks earlier.
To keep up with rapidly changing AI-generated content, LinkedIn uses a teacher-student model. Larger language models identify emerging patterns and generate training data, which is then used to train smaller, faster models that can operate at scale. LinkedIn also uses techniques such as LoRA to update those models without rebuilding them from scratch every time the patterns change.
Humans remain part of the process. LinkedIn’s agentic systems can send difficult cases to expert reviewers, whose decisions are then fed back into the system to improve future classification.
The interesting part is how systematic the effort has become. LinkedIn is no longer relying mainly on user reports or simple AI detection. It is building a continuous learning system designed to adapt as AI-generated content itself evolves.
For creators, that raises the stakes around originality. The question is becoming less about whether AI helped produce a post and more about whether the final content provides enough specificity, expertise, and human value to deserve wider distribution.
Ryan Roslansky Is Leaving Microsoft After Nearly 18 Years
Former LinkedIn CEO Ryan Roslansky is leaving Microsoft after nearly 18 years across LinkedIn and its parent company.
Roslansky joined LinkedIn in 2009 as Jeff Weiner’s first hire and became CEO in 2020. During his tenure, LinkedIn passed one billion members and, according to Roslansky, nearly tripled its business.
More recently, his responsibilities expanded beyond LinkedIn. Microsoft asked him to help lead the future of Microsoft 365, bringing together product, engineering, and design as the company worked to integrate Copilot more deeply into its productivity ecosystem.
Roslansky will remain as an adviser to Satya Nadella through the end of 2026. Dan Shapero will continue as CEO of LinkedIn, while Mohak Shroff remains in his current role, with both reporting directly to Nadella.
Roslansky said the next phase of the Microsoft 365 role would require him to be based in Redmond full time, while his family remains in the Bay Area.
The leadership change comes at an important moment for both companies. Microsoft is trying to build what Nadella describes as a new “OS for work” around Copilot, while LinkedIn is simultaneously embedding AI deeper into hiring, learning, advertising, search, and the feed.
Roslansky helped guide LinkedIn through one of the biggest transformations in its history. His departure now comes just as the platform enters another one.
The Linked Blog is here to help you or your brand have the best possible LinkedIn presence, so feel free to contact us if you need help! See more about what we can do for you here.
