She leads LinkedIn’s AI transformation to create more jobs: ‘I am an AI optimist because I believe in humans’

Imagine this: The role you are applying for was crafted by artificial intelligence (AI). It was AI too, that assessed you to be a good fit for the company – based on the resume you sent, which was also written by AI. As was the message you sent to the hiring manager for your application.

AI as a person’s constant companion in the employment process is not a hypothetical situation for the future. It is already happening now.  

These are some of the new generative AI (GenAI) features recently made available to LinkedIn’s paid users. Driving these changes is its VP of Engineering, and Head of Data and AI, 42-year-old Ya Xu, who leads a global team of almost 1,000 data scientists and AI engineers in creating more professional opportunities for people.

THE FUTURE OF WORK WITH AI AT PLAY

It is hard to imagine that GenAI only entered popular vernacular in November 2022 when ChatGPT, an AI language model, shook the world. It could answer questions relevantly, create meaningful content, and hold human-like conversations.

Over the next 20 months or so, the world watched with wonder and some trepidation as GenAI wrote entire articles, wrote code, and created art, music, videos and even celebrities’ voices and photos.

“The one thing for sure is that in the very near future, the way that we work is going to be different,” said the China-born Ya, who is based at LinkedIn’s headquarters in California, United States.  

The big question then, will AI replace jobs?

“Our chief economist and her team (at the Economic Graph Research Institute) have put out this framework [where] they take all the jobs and break down each job based on the tasks.

“If we’re looking at jobs at the task level, you can say, this task can be enhanced by AI, this task is going to be replaced by AI, and this task is going to be insulated from AI,” she explained.

That is a useful framework to look at LinkedIn’s latest GenAI features, which Ya and her team started to develop in 2022, before ChatGPT was launched.

“LinkedIn being part of Microsoft, we have a front row seat in what the future holds. Several months before ChatGPT was released, I was able to see not just ChatGPT, but GPT-4, a very powerful model that OpenAI wasn’t able to release to the world at that time,” she told CNA Women.

“Sam Altman and Greg Brockman (OpenAI founders) actually came to Redmond in Washington, where Satya (Nadella, Microsoft CEO) essentially hosted a few AI leaders across Microsoft, so I got to see [ChatGPT] ahead of time.”

What she saw propelled LinkedIn’s GenAI technology stack and products. Since November 2023, LinkedIn has been launching the features described earlier to function as a “co-pilot” for their users’ professional life.

The idea is to replace the drudgery of work, Ya, who has a PhD in Statistics from Stanford University, said.

For recruiters, instead of wasting time writing messages to candidates or painstakingly searching for the right candidate, this AI efficiency boost means they can spend more time talking to candidates, to hire the right talent for the role, Ya explained.

And as a candidate, “instead of trying to edit my resume 100 different times, AI can help with that process – instead of spending 10 hours, I only spend an hour”, she added.

As for ethical concerns about AI-written resumes, Ya said: “I think it matters less whether our resume is written by AI, but whether the content is what the human can stand behind. AI can help communication but humans need to be in control.”

A BRAVE NEW WORLD WITH AI

“I know there are AI pessimists – people who think AI is going to take over the world,” Ya said. “I would say, first of all, they think too little of human ingenuity.

“When the automatic teller machine (ATM) first came out, everyone was like, it’s going to take over the bank teller’s job. Fast forward 60 years, you would have imagined that with ATMs everywhere, we would have less bank tellers.

“But no, actually, there are more people working in bank services than ever before. Because all of a sudden, the bank teller’s job is not giving cash, which a machine can do. There is so much more that banks are doing now. Financial management as a career was created,” she said.

“I am an AI optimist because I believe in humans. We are going to create opportunities, jobs for ourselves that today, we cannot imagine because we are so busy doing the drudgery work,” she added.

One area Ya and her team are looking closely at is misinformation. “How can we make sure that when AI speaks with such confidence, we actually have a better guardrail of whether that is hallucination or not?”

Ya also emphasised the importance of fairness, transparency and eliminating bias in an AI-augmented world. When relying on AI to polish a resume for example, this means ensuring that gender bias does not affect how a resume is edited, or the outcomes for the individuals.

“When the first car was invented, I’m sure it was bare metal. There was no seatbelt. The roads were not built to drive the car. There’s no speed bump, no speed limit. That’s how I would want us to imagine AI,” she said.

“Today, we are still early in terms of the power of AI. So you’re driving your car at five miles an hour, and there’s no seat belt, and no speed bump on the road yet. The technology, guardrails and constraints are going to be all emerging.

“And so, years from today, we are going to have a Ferrari that is able to go from zero to 100 miles per hour in a few seconds. And humans still remain in control,” she added.

THE POWER OF AI

Even before GenAI, AI has been one of the most exciting frontiers of technology. LinkedIn began to tap into AI for its “People You May Know” feature around 15 years ago.

Ya joined the company as its first female Principal Staff Engineer 11 years ago, leading the establishment of LinkedIn’s initial experimentation team before assuming her current role as Head of Data and AI in 2021.

“LinkedIn is a matchmaker,” Ya said. It matches talents to jobs, and marketers with buyers in its advertising space. Its feed algorithm decides what trending news or professional updates to put in front of users and in which order, she added.

Ya’s role: To mentor her team, guide the direction and execution of key initiatives, and lead strategic reviews to shape future-forward technology.

She is deeply committed to mentoring and developing women technologists, advocating for LinkedIn's global Women in Tech programme. Among other initiatives, the programme nurtures the interest of young female talents with undergraduate and high school training programmes at LinkedIn.

DEEP LOVE FOR HER MOTHER

Ya was inspired by her own childhood to equal the playing field for women. One of two daughters born in a small village in Sichuan, China, Ya experienced gender bias firsthand.

“Back in the day in China, everyone wanted a boy. Even my grandparents were commenting on the fact that [my mum] was unfortunate to have two girls.

“My mum is a super petite woman who did not have much education. But she did not let anybody define her and her daughters. She believed that we could do better than the boys and would bike us five to 10 miles a day to a better school so that we could have a better education,” Ya recalled.

Ya, who is married with two children, is deeply inspired by her mother’s tenacity.

“When computers started to become a thing, [my mum] learned how to type. Imagine somebody who didn’t even know how to speak proper Mandarin – because we speak the Sichuan dialect – [who had] never seen a computer before and didn’t know how to write much, being able to learn to type in her forties.

“Then, in her fifties, she learned how to swim by pure perseverance. She would go swimming every day.

“That’s what gives me the optimism that you really can do anything if you put your mind to it. As humans, we can do anything,” she said.

Her mother also had a huge influence on her interest in technology.

“When I came to the United States for undergraduate study, there was no such thing as cell phones. It was very hard for [my mum] not to be close to me physically and not to be able to talk to me and see me.

“Fast forward to today, my mum is still in China. We are on WeChat all the time. She can video chat with me whenever she wants to. I call her on my drive home every day. Think about how close technology brought our families.

“I’m a believer that technology is fundamentally helping our lives to be better and to bring this human connectivity more than anything else,” she said.

Her advice for humans in the AI-era: “The space is moving so fast. Having a learner-first mindset is probably the most important when you have a very disruptive technology that can change the future in a way that nobody can predict.”