Amy Lindgren
Second Sunday Series — Editor’s Note: This is the fifth of 12 columns on AI and work, which will appear the second Sunday of each month, from September through August. Last month’s column provided tips on using ChatGPT, while previous columns looked at work opportunities with artificial intelligence; AI use in the hiring process; and an overview of artificial intelligence in general.
Oof. For this column, I had planned to discuss the use of artificial intelligence when developing cover letters and résumés. But then I learned that organizations may be moving even faster than predicted in adapting AI for the workplace.
Change of plans! For this month’s Second Sunday series, I’m frontloading a topic originally slated for later in the year. Today I’m discussing AI best practices for employers and managers. If you don’t fit in that category, read on anyway. As things are going, your current or future employer might be part of this growing trend.
AI use in hiring
While statistics for fast-moving situations can be tricky, it seems clear that most of the largest organizations — think Fortune 1,000 — are using AI-enhanced applicant tracking systems (ATS), as well as other AI tools when hiring.
Smaller organizations are more difficult to track but survey respondents indicate high AI usage as well. In an older (2023) survey by ResumeBuilder.com, 43% of responding hiring managers confirmed current or upcoming use of AI interviews, with 15% planning to use AI to make hiring decisions without any human input.
Pause to consider that. Even if actual numbers vary, knowing that a countable number of employers intend to let AI hire their workers means we’ve entered a new world of … everything. Including error. Case in point: in the same survey, 79% responded that it’s very or somewhat likely that AI interviews screen out viable candidates more frequently than human interviewers do.
Best Practices: To start, plan AI usage to treat candidates the way you would want to be treated. Also, if you’re increasing automation because too many candidates apply, rethink your hiring practices to cast a less-broad net.
And if you worry that casting the net less broadly means a non-inclusive process, remember that AI routinely disqualifies candidates through errors in its own data. For example, AI interviews can wrongly interpret speech impediments or candidate disabilities.
As you look for appropriate models for AI hiring, don’t forget about government resources. For example, this link leads to a tool recently released by the U.S. Department of Labor called “The AI & Inclusive Hiring Framework”: https://www.peatworks.org/ai-inclusive-hiring-framework/
AI use for work processes
In the span of just a few years, AI products and services for workplace use have grown exponentially. As a result, organizations are lagging when it comes to policies and standardization. Indeed, it’s common for workers at all levels to use AI tools they’ve discovered independently without even telling their supervisor.
Best Practices: Ideally, organizations would create systemwide AI policies based on analysis and research. Instead of generally assuming AI tools will create efficiencies, each use would be vetted and measured before being incorporated — particularly if job loss would result.
Related to staffing and AI, best practices include shifting the focus from the tools to the team using them. Investing in staff training (rather than using the fire-rehire model) ensures that institutional knowledge informs use of the tool, while motivating team members with paid training or bonuses for completed certifications creates buy-in for the change.
Best practices would also demand standardization on key points such as transparency and data privacy.
Reliance on AI accuracy
As a culture, we seem inclined to believe our digital tools. If spell-check doesn’t understand a word, we let ourselves be bullied into using another word instead. This example is small, but it should remind us that artificial intelligence can and does produce false results.
Best Practices: Whenever an organization introduces a new AI use, the smart practice would be to simultaneously implement an evaluation or testing process to ensure it’s solving more problems than it’s creating. To be effective, evaluation criteria has to be based on more than just speed — because making errors the fastest isn’t exactly a win.
As always, there’s so much more that can be said about this AI topic. But we should return to our regularly planned program. Come back next month and we’ll jump in with ways to use AI as a job seeker, including tools to help with resumes and cover letters.
Related Articles
Working Strategies: Four resolutions for 2025
Working Strategies: Starting fresh on job search in the New Year
Working Strategies: When the holidays stall your job search
Working Strategies: A reading list for winter break
Working Strategies: Second Sunday Series: Jumping in with ChatGPT
Amy Lindgren owns a career consulting firm in St. Paul. She can be reached at alindgren@prototypecareerservice.com.
Leave a Reply