Recent research has indicated that 99% of Fortune 500 companies are using some kind of AI, or automated tool to screen and rank candidates for hire. What’s this doing to our older workers? [Transcript 1 | Transcript 2 | Video 1 | Video 2]

5 takeaways:

➀ “Artificial Intelligence can exacerbate bias and make it worse,” AARP’s Director of Finance and Employment Lori Trawinski told Living Longer fellows. Amazon built a resume sorting algorithm that was using a biased data set and ended up not pointing to any women as hirable. “We often hear the story that there are not as many women programmers as there are men. There are not as many people of diverse backgrounds, different races and ethnicities, sexual orientation, gender” Trawinski said. Creating a programming group that has diversity can help lead to a less biased model. AI can filter through graduation dates, gaps in working and more – all affecting the hiring of older workers.

➁ There are four different types of artificial intelligence:

1. Machine learning provides computers with the ability to learn things, Trawinski said. The computer will take in a large data set and will attempt to look for patterns in the data.
2. Natural language processing is when the computer looks at written words or verbal communications to learn from the language that it’s looking at.
3. Deep learning is a series of machine learnings that build upon one another. A mammogram is an example of deep learning, Trawinski said.
4. Neural networks “is where the systems are patterned after the neurons, attempting to replicate how the human brain works,” she said. This is used in facial recognition technology or handwriting recognition. The documentary “Coded Bias” provides some excellent examples of how governmental agencies in different countries use facial recognition technology to do things that average people aren’t aware is happening, Trawinski said.

➂ There are big gaps between how older and younger adults view Artificial Intelligence. Chaiwoo Lee from the MIT AgeLab studied the intersection of the increased use of AI and increased longevity, including financial services and retirement planning, health and caregiving, information and community, and information and communication. “We’re seeing some big gaps in terms of how older and younger adults differ in terms of their confidence and knowledge,” Lee said. Adults are uncertain about its risks, and experts have overestimated consumer willingness to use AI in all kinds of different contexts, she said.

➃ The Equal Employment Opportunity Office is trying to make AI-related bias clear for employers, employees and investigators. The EEOC’s initiative on Artificial Intelligence and Algorithmic Fairness released two technical assistance documents in May, Associate Legal Counsel Raymond Peeler said. One is the Americans with Disabilities Act and The Use of Software Algorithms and Artificial Intelligence to Assess Job Applicants and Employees and the other is Tips for Workers on Disability Discrimination and the Use of Software. The EEOC is “developing concrete suggestions to help employers and employees reap the potential benefits of AI while still complying with civil rights laws,” and is looking internally to create staff training, to ensure that investigators have tools needed to identify and address AI-related bias when it may arise during a charge investigation, Peeler said.

➄ Journalists should ask, “How many people know that HR systems are sorting resumes like this? Also, “Are people getting training on this? Trawinski asked. “There are some elite schools that train people about how to apply for jobs. If you don’t go to an elite school or you’re an older worker, how do you know?” Other questions journalists should consider include:

• Are people with disabilities, like autism, being treated with equity during the interview process?
• Is arthritis in older workers causing them to be screened out due to their lack of keyboard or computer skills?
• As job advertisements are increasingly being promoted on social media accounts, are older individuals being excluded?

This program is sponsored by AARP. NPF is solely responsible for the content.