California’s SB 7 is part of a larger trend in AI regulation as a national lawmaker …(+)
Getty
While checking criminal history has long been an important employment tool, access to criminal history has become increasingly difficult.
In California, a 2021 court ruling has made it more difficult for background screening companies to verify criminal history and provide less ways for employers to assess candidate risk. As a result, some companies are relying on AI-powered social media screening to fill the gap. They use automated systems to flag potential risks based on public online activities.
But California Senate Bill 7 (SB 7) could quickly change that. If enacted, the Act introduces new regulations regarding the Automatic Decision System (ADS) in Employment. The bill would prohibit employers from using advertising to infer criminal or credit history, including AI-powered social media screening tools.
For employers who rely on background checks to make informed employment decisions, the law could affect current screening practices. With fewer tools available to assess risk, hiring a team may need to reassess the approach while ensuring the evolution of AI regulations.
What is an automatic decision system?
SB 7 defines an automated decision-making system (ADS) as a computational process using machine learning, statistical modeling, data analysis, or artificial intelligence to use machine learning, statistical modeling, data analysis, or artificial intelligence to assist or replace human decision-making and generate simplified output that affects individuals.
In particular, the bill excludes certain common workplace technologies from this definition, such as spam email filters, firewalls, antivirus software, and identity management tools. This distinction suggests that the bill is not a complete ban on AI, but an effort to regulate how automated systems affect employment decisions, particularly when exchanging or supplementing human discretion.
New Compliance Requirements for Employers
If passed, SB 7 introduces a major new obligation for employers to use advertising in employment or employment decisions.
First, the bill requires greater transparency. Employers using advertising must notify workers in writing at least 30 days before deploying the system. Notifications should explain how the system works, what kind of data it uses, and how it affects employment decisions. Employers should also maintain an updated list of all advertising tools they are using.
SB 7 goes beyond disclosure to limit how AI can be used in hiring decisions. Employers are prohibited from using advertising to infer sensitive characteristics such as criminal history, immigrant status, gender identity, political beliefs, and creditworthiness. This means that AI systems cannot analyze indirect data, such as social media activities, employment gaps, and financial behavior, to make assumptions about candidates’ backgrounds.
This bill also requires human surveillance. Employers cannot rely solely on AI-driven output. When advertising is used, the human reviewer must independently evaluate the system’s conclusions. This introduces a new layer of compliance and requires businesses to balance efficiency with fairness in their employment decisions.
How SB 7 affects employers doing background checks
Employers already face a major hurdle in examining criminal history, and SB 7 could make this even more challenging.
We all or no v. Hamrick has fundamentally changed the way background check providers verify their employer’s criminal history. The ruling prohibits California court records from displaying key identifiers such as date of birth and driver’s license number. This has made it much more difficult to see if the record actually belongs to a candidate.
Though it is aimed at protecting privacy, this decision has posed a major validation challenge. Employers require background checks, especially for candidates with common names, but screening companies can’t confidently decide whether the record belongs to the applicant in question. Eliminating the identifier increases the likelihood of both false positives and false negatives, and asks employers to assume the risk adoption or to seek alternative screening methods.
The role of AI in filling the gap and why SB 7 challenges it
Due to limited access to verifiable criminal history, some employers have turned to social media screening powered by AI as an alternative risk assessment tool. These tools scan publicly available online content and flag potential risks related to categories such as violence, drug use, and financial irresponsibility.
However, SB 7 is directly challenging this practice. Employers will no longer be allowed to use advertising tools that infer offenders or credit history from social media activities, even if the system does not expressly state that it is making such inferences. The ban may also apply when AI uses proxy metrics, such as discussions of past legal troubles, partnering with certain groups, or posting about financial difficulties.
This has implications for employers. Companies that previously relied on AI-driven insights to compensate for the gaps in criminal history screening may quickly feel that there is no legal basis to continue using such tools. Employers who have built AI-powered risk mitigation strategies should rethink their approach to avoid compliance pitfalls.
The future of AI in employment decisions
California’s SB 7 is part of a larger trend in AI regulations as lawmakers across the country try to prevent algorithm bias and increase job transparency. However, for employers, the law could present more challenges than opportunities.
On the one hand, employment costs and decision-making can be slower, requiring greater surveillance at multiple stages. Meanwhile, it aims to promote equity, but could add another layer of regulatory complexity and limit the employer’s ability to effectively assess candidate risk. The intention is to reduce bias, but we have yet to see the impact of limiting AI tools on adopting fairness and efficiency.
To prepare for these changes, employers should:
Audit AI-driven employment tools to determine whether they rely on estimated characteristics rather than verifiable data. Review the role of AI in social media screening and assess whether adjustments to flagged content categories can reduce compliance risk. Develop a compliance-first AI strategy that balances transparency, risk mitigation and legal requirements.
While AI continues to be an important part of the hiring process, SB 7 may show a change in the way employers leverage technology for screening. As regulations evolve, businesses must find ways to balance risk mitigation and compliance. Employment decisions need to ensure that they remain effective and legally defensible in increasingly complex landscapes.
