HR has already seen the boons and bumps of artificial intelligence creeping its way into the day-to-day workflow. Beyond simple predictive text in document processors, Google unveiled AI integrations for Docs in March; the updated Google Workspace can now generate bare bones job descriptions and revamp the tone of emails.
The tech company also unveiled GenAI integrations in UKG software in May. That month, Microsoft also built on its previous offerings — namely, Microsoft 365 Copilot — and published a trends report titled "Will AI Fix Work?" LinkedIn Recruiter also launched AI-fueled outreach messages in Q2 2023.
And of course, a certain problematic chatbot is on the tip of everyone’s tongue these days. (Although, Glassdoor estimated that ChatGPT supporters or sympathizers outnumber the naysayers in an 80-20 split.) But many employers and HR still don’t feel well-versed in all that artificial intelligence and machine learning, known as AI/ML, can do or has to offer.
A crash course in AI offerings for HR
While a bit heavy with jargon, the Josh Bersin Company recently released an informative guide to different types of AI for work. The three categories the report focused on are emerging AI, first generation AI and second generation AI. “Emerging” software features are “added on,” “first-generation” features are built in, and “second-generation” describes software solutions that are built on the work of AI.
Researchers said they created the resource to “understand exactly what's preventing HR from fully exploiting artificial intelligence in core business strategies,” including recruiting and talent management.
Examples of emerging AI that HR pros may already use included technology like predictive analytics, language processing, intelligent chat and image generation. (Think Midjourney, DALL-E or even Photoshop’s AI integration.) The next step up or first-generation features include machine learning or advanced candidate matching. Researchers gave Workday, LinkedIn and SAP as examples.
And finally, the heavy hitter: second-generation AI. A lot of the technology explored in this section of the report was complex (vector bases, large language models and what the Josh Bersin Company called “advanced models”). That being said, an interesting concept for HR pros to watch are neural networks. The purpose is in the name: “A neural network is an algorithm that mimics the human brain by having individual, simple units called ‘neurons’ that take in a bunch of inputs and compute a simple function to create an output,” researchers said.
In turn, “deep learning neural networks” involve a high level of layers and interconnections, researchers continued, made possible by technological advancements such as the increase in what computers can process and the declining price of access to such computation.
In short, as technology advances, the future of what’s possible with AI at work expands.
The pitfalls of AI
Because so much of traditional AI is based on human input, the pushback against this type of technology has historically focused on how human bias informs the use of such tools.
Last year, New York City effectively restricted AI in hiring. Later, the U.S. Equal Employment Opportunity Commission’s guidance would echo NYC’s call to action, which demanded that employers audit their automated decision-making tools. In a public hearing at the top of the year, EEOC Vice Chair Jocelyn Samuels announced the agency’s intention to ensure that AI doesn’t replicate discriminatory human approaches to hiring.
Is AI a diverse talent pipeline solution?
A few months later, a tech job board Dice.com agreed to use AI to root out national origin bias in its job listings as a part of a settlement with EEOC. Likewise, Josh Bersin Company researchers nodded to major hiring discrimination suits settled against Google and Goldman Sachs.
Reaffirming the utility of practices such as that of Dice.com’s EEOC settlement, researchers wrote that second-generation AI could help with recommending job opportunities “independent of gender” and could ultimately help with promotions.
“If you added your company’s pay data to this system, you could certainly see the system giving you ‘predictions’ for pay, which would immediately show groups who are underpaid, overpaid, or otherwise,” researchers said. “Pay equity and diversity and inclusive promotion are enormous new applications for AI.”