What Counts as an "AI Tool" Under Illinois HB 3773

The single biggest misread of HB 3773 is the definition of "AI tool." Employers picture shiny dedicated AI products and conclude they're out of scope. Most are wrong. The law reaches a much wider set of features — many of which are quietly embedded in tools you already own.

Here's a practical taxonomy of what counts, with examples of the categories companies routinely miss.

The legal framing

HB 3773 defines AI broadly — it isn't limited to bespoke "AI products." Anything that uses algorithmic decision-making, pattern recognition, or automated scoring to influence an employment decision is in scope. That definition is intentionally functional, not technological, because the harm the statute targets is discriminatory outcome, not vendor branding.

Six categories that trigger obligations

1. Applicant tracking systems with matching, ranking, or scoring

Most modern ATSs added some form of AI matching in the last 24–36 months. If your ATS shows you a "fit score" or reorders candidates by predicted relevance, it's in scope. This includes features that are off by default but available to your recruiters.

2. Video-interview platforms that score candidates

Video tools that evaluate facial expression, speech, or word choice to generate scores or recommendations are clearly within scope. Even tools that only transcribe video may be in scope when paired with downstream AI summarization.

3. Résumé screeners and AI-enabled sourcing

This category includes both dedicated screening tools and the AI features inside platforms like LinkedIn Recruiter — relevance ranking, candidate suggestions, and message personalization based on candidate profile signals.

4. Generative-AI assistants (ChatGPT, Copilot, Gemini)

This is the category companies miss most often. When an HR coordinator pastes a stack of résumés into ChatGPT and asks for a ranking, an "AI tool" is in use — even though the company never bought "an AI screening product." The same is true when a hiring manager uses Copilot to summarize interview notes that then feed a hiring decision.

5. Assessment and skills-testing platforms

Algorithmic scoring of cognitive tests, situational judgment exercises, and skills simulations is in scope when the score is used in an employment decision.

6. Performance management and scheduling tools

Predictive scheduling, attrition-risk scoring, and AI-assisted performance review summaries are all in scope. This is the most-overlooked category because companies don't think of their workforce-management software as "AI."

The four blind spots

1. Features that were quietly added to tools you already owned (the ATS matcher you didn't ask for)
2. Informal generative-AI use by your team — pasting résumés or interview notes into a chat assistant
3. Tools used by hiring managers outside HR's visibility (sourcing extensions, AI plugins for email)
4. Vendor-side AI running inside the platforms you use for benefits administration, time and attendance, or learning management

The reasonable test

If you're not sure whether a feature counts, ask: "Does this tool influence an employment decision in a way a candidate or employee would want to know about?" If yes, assume the statute reaches it and inventory accordingly.

What this means operationally

An honest inventory will almost always find more AI than leadership expected. That's why "no complete AI inventory" is gap #1 in most companies, and why we start every engagement there. The deeper take on scope is in our plain-English pillar guide.

Lakeshore is a compliance consultancy, not a law firm. This article is general information and not legal advice.