AI

AI @ LinkedIn - It’s All About Foundations

For our engineering teams, artificial intelligence (AI) is like oxygen - it powers every product we build and every experience we deliver. Our members can see it in things like the job recommendations they’re served, the content that’s personalized for them in their feed, and the LinkedIn Learning courses that are tailored to their interests. Our customers benefit from it through more relevant recommendations in LinkedIn Recruiter and better ways to engage with sales prospects in Sales Navigator.

At the same time, our approach to using AI is the same approach we take to any other new tool or technology, we start by going back to our mission and vision. This guides us in the tools and technologies we build, with a focus on delivering value to our members and customers. We think about how AI should help everyone at every level in every organization do better in their job and in their career – like connecting with each other, helping them find the right jobs, and gaining the right skills.

AI and the Economic Graph

Very early on, we realized the importance of converting LinkedIn’s Economic Graph, which is a digital representation of the global economy, to a data model. We did this to bring together information about members, companies, jobs, schools, and skills to help us all better understand the rapidly changing world of work through this detailed and dynamic data. 

As the Economic Graph evolved, we were able to gain richer insights and apply those insights in new ways. We used machine learning to help us build out macroeconomic models to help us forecast government labor data and to understand how the skills associated with certain occupations are changing over time. AI is one of our core differentiators, and it allowed us to take this data set and drive actionable insights, analysis, and visibility from it, while doing so in a trusted manner that preserved our members’ privacy.

With Great Power, Comes Great Responsibility

The responsible use of AI has been fundamental to our efforts for years. Two years ago we open sourced the LinkedIn Fairness Toolkit because we recognized the importance of helping people measure fairness and bias, as well as detecting statistically significant differences in model performance across different subgroups.

Our work with AI is not just about what we deliver to our members and our customers, but how it's delivered to them. With every initiative, fairness, equity, and responsibility get built into the core of our work. But we know we can keep getting better, so we doubled down and codified our best practices into a single set of Responsible AI Principles. Built in alignment with Microsoft’s Responsible AI Standard, these principles guide all our AI work across LinkedIn and enable us to advance economic opportunity, uphold trust, promote fairness and inclusion, provide transparency, and embrace accountability.

The Road Ahead

New AI approaches, like generative AI and Large Language Models (LLMs), have already shown their potential to help our members and customers become more productive and successful. When we bring together generative AI with datasets like our Economic Graph, that’s where we can get even better at helping professionals connect to opportunities, showcase their skills, and gain the knowledge they need to be better at their job and in their career. Almost 20 years ago this company was founded with a clear vision – create economic opportunity for every member of the global workforce – and we’ll continue to use AI responsibly to help us accelerate our progress toward that vision.