Job searches and generative AI seem to be a good match. The large-language models used for such tools as ChatGPT and Google’s Bard can help job seekers in everything from writing resumes and cover letters to surfacing the employment opportunities that best fit their interests and skills.
And it seems to be working. According to a study in February by job-seeking site ResumeBuilder, 46% of those looking for jobs used OpenAI’s ChatGPT to write resumes and cover letters – and that was only three months after the tool was launched in November 2022 – and seven in 10 who did use ChatGPT reported a higher response rate from employers.
In addition, 78% got an interview and 59% were hired.
LinkedIn for more than a decade has used AI technologies to connect people to jobs and that has only accelerated with the rapid ascent of generative AI, which isn’t surprising given the massive job site’s parent, Microsoft, has invested more than $10 billion in OpenAI and is aggressively integrating AI throughout its product and services portfolio.
In a blog post in March, Tomer Cohen, LinkedIn’s chief product officer, said the company’s goal is always to help more people worldwide solve various professionals needs at every stage of their careers, and “with advanced generative AI and large language models, we’re now able to do that. And we do so by putting AI directly into the hands of our members.”
New Features for Premium Users
That has included an AI-powered profile writing and messaging features. This week – as the number of members surged past 1 billion – LinkedIn is adding to its generative AI-based services with new features for Premium users, who pay $39.99 a month. (The company also has a free membership level.)
Starting with a select group of Premium subscribers, the new AI features, based on OpenAI’s GPT-4 LLM, will bring more personalization capabilities and help users sort through the mountains of information put before them. That includes analyzing posts like articles and conversations in a user’s feed and scouring long articles and videos to pull out information most relevant to the subscriber.
The technology also uses Microsoft’s Bing search engine to quickly deliver needed information, Cohen wrote this week.
To pull off this level of personalization, LinkedIn engineers combined OpenAI models with the company’s internal data, a technique more organizations are doing to create generative AI tools that more closely fit their needs.
“We not only built our own tooling and platforms to utilize the capabilities of existing OpenAI models in our tech stack, but also leveraged the knowledge we have about our members’ professional skills and interests, as well as insights across LinkedIn, in order to create a personalized experience for our members in a unique, LinkedIn way,” Lei Yang, vice president of engineering at LinkedIn, wrote in a blog post.
They also enabled the model to provide personalized responses within about five seconds by reaching a balance between personalization and response speed, including building mechanisms like streaming responses, offline data processing, and pre-computation to reduce latency, Lei wrote.
Shoring Up the Weaknesses
The engineers also took steps to address challenges that arise with generative AI chatbots and other tools, such as bad or incorrect responses, aka “hallucinations.” That included building tools and processes to evaluate response quality and hallucination rates, improving the engineering around prompts. During evaluations, the tools also analyze the tone of responses to ensure they’re professional.
“Finally, the team also adds trust checks to establish safety boundaries and performs thorough red teaming trust exercises regularly,” Lei wrote.
This is important as generative AI chatbots take on larger roles in interacting with humans, including those making hiring decisions. A survey this year by recruiting-maker iCIMS found that 47% of recent college graduates were interested in using ChatGPT or other chatbots for resume and cover-letter writing and that 25% already had done so.
“Generative AI bots (like ChatGPT) have the potential to help both candidates and employers be more productive and communicate better,” iCIMS CTO Al Smith wrote in the survey. “From resumes and job descriptions to offer letters and onboarding material, the technology offers employers and recruiters a better starting point as they move through the hiring journey.”
Job Seekers Beware
Still, those angling for jobs need to be cautious, according to Robert Henderson, a career advice writer and resume expert with job-seeking site Jobscan. ChatGPT will help them save time and money streamlining resume writing and other tasks, Henderson wrote in a blog post.
However, the chatbot has its shortfalls, from hallucinations and not identifying important keywords to a lack of personalization, which includes not highlighting a person’s particular strengths or using language that makes the resume read like everyone else’s.
It also might prevent a person from getting a job. Pointing to the ResumeBuilder survey, Henderson noted that 40% of respondents said the interviewer knew they had used ChatGPT to write the resume and cover letter and that, of those, 35% said they didn’t get the job because of it.
That echoes iCIMS’ survey, which found that 39% of HR professionals said the use of AI during the hiring process would sink the job seeker’s chances.