By: Sarah Johnson
This week we’ll take a closer look at some very intriguing legislation, Illinois HB 2557, the Illinois Artificial Intelligence Video Interview Act. This bill, which had passed both chambers by the end of May 2019, demonstrates the role artificial intelligence (AI) is starting to play in our society. I find AI to be an incredibly fascinating technology and am constantly interested as we discover more ways it can be used throughout our lives. This bill is a great example of an application I hadn’t really considered before. So let’s take a closer a look at how the Illinois legislature is attempting to regulate AI’s role in the professional hiring process.
Background on AI
First we need to understand what AI actually is. AI is not actually defined in the bill, but is traditionally considered to be “the simulation of human intelligence processes by machines, especially computer systems. These processes include learning (the acquisition of information and rules for using the information), reasoning (using rules to reach approximate or definite conclusions) and self-correction.” Essentially, AI is the capability of a machine to imitate intelligent human behavior.
There are two different categories of AI, weak and strong. Weak, or “narrow”, AI is designed and trained for a very particular task. The most familiar type of weak AI is virtual personal assistants, think “Hey Siri”, “Alexa ___” or “Okay Google”. Strong, or artificial general intelligence, is an AI system that works more like a human brain, clustering and associating data similar to human cognitive abilities. Think, you telling a connected machine “Good Morning” and then turn on the coffee maker – the machine will start to associate the two and eventually perform this action for you through “learning”. The AI system is able to find solutions when presented with an unfamiliar task. The robots in I, Robot are familiar, more advanced, example of strong AI.
The Bill
HB 2557 implements three major regulations on employers looking to use AI at any point in their hiring process. The bill prohibits Illinois employers from using “artificial intelligence” as a means to evaluate job applicant’s video interview and consider the applicant’s fitness for the position unless the employer:
- Notifies the applicant
- Explains “how the artificial intelligence works”
- Obtains (before the interview) consent to evaluation by the artificial intelligence program
There are other restrictions on the employer and how they conduct their interviews with AI. After a video is completed and the AI is used, employers are restricted from sharing video interviews and required to delete the videos within 30 days of a request by the potential employee.
Discussion points
There are a few things I find interesting about the bill. First, they do not define AI in the bill. There are generally accepted definitions of AI, but it is not a self-evident term. The extent to which it is defined is important because we need to know what is subject to be included in the provisions the bill sets forth. How far does the technology have to go to qualify as AI under the bill? Is weak AI included? Or just strong? Somewhere in the middle?
This then brings me to wonder what methods employers will use explain “how the artificial intelligence works” and the “general types of characteristics” the AI is using in its evaluation. How in depth must employers go? AI can get pretty technical, what is the level of information that must be communicated to the potential employee to explain the how? Of course, this depends what the company uses that is “AI”, ranging from judging an applicant’s behavior (facial expressions, tone, quality of an answer – or strong AI) to making a video more digestible for human reviewers by automatically editing out dead parts (weak AI).
Another concern is the possibility of AI perpetuating certain biases, examples of which have already been seen. AI is usually “trained” by comparing data, in this case incoming interviews, with success, in this case who actually got hired by the company. Unfortunately, if a company has traditionally skewed towards (or away from) certain categories of people, the AI will learn to do the same unless the training is handled very carefully to avoid this outcome. If knowledge that AI is being used causes certain groups to withdraw their application or request not to be evaluated with this method, would that only exacerbate this potential problem?
Ensuring all backups of videos are deleted within 30 days upon a participant’s request is also potentially problematic. Do participants have to be informed their content may not be deleted if they do not request it to be? What if the company uses a third party to conduct and/or process videos? Are they subjected to the same restrictions? How will they ensure all copies are deleted? What happens if it turns out not all copies were deleted?
What do you think about this bill? Have you seen any other similar legislation? What do you think about AI starting to play a more dominant role in our society?
Cover Photo by Alex Knight on Unsplash
About BillTrack50 – BillTrack50 offers free tools for citizens to easily research legislators and bills across all 50 states and Congress. BillTrack50 also offers professional tools to help organizations with ongoing legislative and regulatory tracking, as well as easy ways to share information both internally and with the public.