LinkedIn taps GPT-4 to polish profiles and write job ads

Writing a LinkedIn summary has always been an awkward exercise in self-promotion. You want to sound impressive without sounding like a corporate robot. Now, Microsoft-owned LinkedIn wants to let actual robots do the work for you.

LinkedIn just rolled out AI-powered writing assistants for personal profiles and job postings. Interestingly, the company is splitting its model stack based on the use case: it is tapping GPT-4 to spruce up user profiles, while using GPT-3.5 to help recruiters draft job descriptions.

GPT-4 for you, GPT-3.5 for job posts

The profile writing tool is rolling out first to paying LinkedIn Premium subscribers. It is aimed at users who know what they have done in their careers, but struggle to package it into an enticing narrative.

You provide some raw bullet points about your work history, and the system drafts a polished story for your About and Headline sections. LinkedIn claims that using the tool helps you stand out for almost twice as many opportunities. It is not clear whether that boost comes from superior algorithmic optimization or simply LinkedIn giving extra visibility to users who adopt the feature.

Recruiters get a similar shortcut. An HR manager can input basic details like a job title and company name, and the tool spins up a full job description to edit and publish. Alongside these product updates, LinkedIn also spotlighted 100 existing AI courses on LinkedIn Learning while adding 20 new courses specifically focused on generative AI.

The ghost in the profile

While drafting text in seconds sounds convenient, it introduces an immediate authenticity issue.

LinkedIn confirmed that it has no plans to disclose or label when profile copy was generated by AI. That means recruiters browsing candidates have no way of knowing whether a sharp personal statement reflects someone's genuine communication skills, or if they simply clicked an AI button.

While job postings are already heavily templated and nobody expects them to convey personal charm, candidate profiles are meant to represent individual credibility. Masking real communication ability behind a generic GPT-4 sheen could end up wasting more time for recruiters trying to get a real read on candidates.

Why it matters

For everyday professionals, having an automated copywriter lowers the barrier to building a presentable profile, especially for non-native speakers or people who dread talking about themselves. That said, it hardly justifies paying for a Premium subscription on its own, since you can easily replicate the same result in any standard ChatGPT window.

Strategically, LinkedIn's model allocation is clever. It locks the expensive GPT-4 behind its consumer paywall to drive high-margin Premium conversions, while assigning the cheaper GPT-3.5 to recruiters who generate bulk job ads at scale. It is a textbook example of matching model intelligence and inference cost to user monetization tiers.

The real danger is platform-wide blandness. Without clear labels, LinkedIn risks flooding its network with sanitized, homogenized executive bios. When job seekers use AI to write their profiles and recruiters use AI to draft and filter job postings, recruitment becomes an automated loop where algorithms talk to algorithms.