LinkedIn news
& changelog.

LinkedIn has introduced a Post Proofreader for Premium members, replacing the AI post enhancement feature it removed the month before. The tool does not rewrite posts but suggests corrections to an existing draft, with changes marked inline so users can accept or reject them individually. The move follows criticism of AI-generated content, after LinkedIn received more than a million reports through its new AI slop button within three weeks.

In its EU transparency report under the Digital Services Act, LinkedIn reports a 46% increase in detected instances of inauthentic activity in the first half of 2026 compared with the second half of 2025. Named categories include engagement pods, external apps offering automated posting, and fake profiles. For users of third-party tools this means automation that manufactures engagement is increasingly being detected and acted on.

Version 202609 of the LinkedIn Marketing API brings four changes: the Ads Reporting API supports a new MEMBER_DESIGNATED_MARKET_AREA pivot for reports grouped by Nielsen DMA regions, the Conversions API accepts the new hashedFirstName and hashedLastName fields (SHA-256) and allows a 180-day attribution window, and the Company Intelligence API gains campaignGroup, objectiveType and seniority filters. Versions 202510 and 202511 sunset on 15 October and 16 November 2026 respectively.

LinkedIn says more than one million members used the "seems like AI slop" report in its first two weeks. Content flagged as slop now gets around 40% fewer views than before the feature launched. Chief Product Officer Hari Srinivasan stresses that a single report does not automatically cut a post's reach; it primarily affects the feed of the member who reported it. Authors who collect many reports will start receiving a notice.

LinkedIn now suggests clips and chapter markers automatically for recordings of LinkedIn Live events. The AI marks moments such as announcements, key insights and quotes; hosts can extend, trim or discard each suggested segment in the editor. Chapters are meant to make long replays easier to skim. LinkedIn says the feature is available globally.

LinkedIn published a playbook on how brands should structure content to be cited by AI chatbots, framed as the "Credibility Stack". Its main points: long-form articles are cited more often than feed posts, content freshness matters, and visibility improves when the company page, employee profiles and creators publish on the same topic. LinkedIn cites third-party research ranking the platform among the most-quoted sources in AI answers.

LinkedIn no longer shows one fixed comment order to everyone: comments under a post are now ranked by relevance to each individual reader. LinkedIn cites professional interests, connections and past engagement as signals. It is also surfacing more recent discussions in the feed. The context is LinkedIn's Q2 report, which showed an 18% year-over-year increase in time spent in post comments.

With API version 202608, the Matched Audiences API is generally available — developers can now apply for access directly through the Developer Portal instead of relying on a limited-access program. At the same time LinkedIn introduces a cap of 1,000 DMP segments per sponsored account; requests beyond that are rejected with error code SEGMENT_LIMIT_EXCEEDED (HTTP 429). The release also adds two new conversion types (MARKETING_QUALIFIED_LEAD, SALES_QUALIFIED_LEAD) and account-level UTM tracking parameters.

LinkedIn ends support for legacy geo targeting in Ads on August 31, 2026. Campaigns still created with legacy geo after that date fail with an HTTP 400 error and the code INVALID_VALUE_FOR_FIELD; all partners must move to Bing-geo. The "location" field is also removed from /v2/me and /v2/people responses — affected integrations need to migrate to the "geoLocation" field with the geo~ projection.

Members can now flag posts as "seems like AI slop" from the post menu. According to LinkedIn, those reports feed classifiers that reduce AI-generated content in out-of-network recommendations. LinkedIn is also replacing its "Enhance your post" AI writing feature with a proofreading tool that leaves the author's voice intact; Chief Product Officer Hari Srinivasan calls AI spam a top priority.

An analysis by detection vendor Pangram finds that 41% of long-form posts on LinkedIn show signs of AI generation — the highest share among the platforms measured (X: roughly one third). LinkedIn says it applies reach penalties when AI spam is detected. For content tools this means generated drafts still need real editing.

LinkedIn adds two metrics to post analytics: reach inside and outside your own network. For the first time you can natively see whether content travels beyond your connections, previously this took third-party analytics tools to approximate.

LinkedIn renames its quick-apply flow and rolls out a Premium Apply Assistant that helps users complete applications. More AI assistance moves directly into the application process, relevant for job seekers who previously relied on external tools.

LinkedIn expands Campaign Manager with AI features: auto-generated ad copy, a Brand Kit and multiple ad variations with different headlines from one draft. B2B marketers can test campaigns faster, generated copy still deserves an editorial review before going live.

The profile skills section now supports 'Connected Apps': users can back up skills through integrations with supported apps and learning platforms. Verified skills are likely to gain weight in recruiter search, relevant for job seekers and sourcers alike.

LinkedIn is testing additional topic-based feeds shown alongside the regular main feed. They are suggested from the topics a user engages with in the app, or from generally trending news stories. The test initially ran with a selected group of users around the Cannes Lions festival; the main feed is not replaced.

The assistant handles boolean sourcing, drafts outreach and schedules interviews, directly inside the Recruiter seat. Rolling out gradually since October 2024.