Across the US, regulating automated employment decision tools is starting to become more of a priority. One state that has taken steps to increase the accountability surrounding the use of technology in the workplace and reduce the potential harm is California, which has proposed California Workplace Technology Accountability Act (AB-1651). The main contributions of this Act are to restrict the data that can be collected about workers to only activities that have proven business necessity, give workers access to their data, and require data protection and algorithmic impact assessments of worker information systems and automated decision tools, respectively.
Here are the 10 things that you need to know about this proposed California Workplace Technology Act:
An automated decision system (ADS) or algorithm is a computational process, including those derived from machine learning, statistics, or other data processing or artificial intelligence techniques, that makes or assists with making employment-related decisions. The output of these systems is any information, data, assumptions, predictions, scoring, recommendations, decisions, or conclusions generated by an ADS.
Workers can request information about the information an employer is collecting, storing, analysing, interpreting, or disseminating about them, including:
Workers also have the right to review their data for inaccuracies and request corrections.
Employers that control data collection shall inform workers regarding the categories of data to be collected and explain how the data collection relates to essential job functions. Furthermore, workers must be informed when employers collect data to make or assist employment-related decisions. Employers also need to answer relevant questions, such as whether data will be deidentified and whether the employers will use the data at the individual or aggregated level. In addition, the question of whether data is being shared with a third party, who they are, and why data is being shared will need to be addressed. Further notification requirements include the length of data retention, workers’ right to access and correct data, relevant data impact assessments, and any active investigations by the labor agency.
Electronic monitoring of workers or collection, storage, analysis, or interpretation is only permittable if allowing a worker to accomplish an objectively proven job function. In addition, it is also allowed when used to monitor production processes or quality. Finally, electronic monitoring of workers is often applied to ensure compliance with employment laws, protect workers' health, safety, or security, and administer wages and benefits.
Employers or vendors acting on behalf of an employer that plans to electronically monitor workers should give a clear and conspicuous notice of their planned activity and inform them of their right to correct their data. This should include:
Within 30 days of the legislation going into effect, employers or vendors acting on their behalf must provide notice to workers of the use of an automated decision tool through the routine communication channel. Notices should outline the nature, purpose and scope of decisions that will be influenced by an ADS, the types of ADS outputs, and the specific category and sources of worker data that the system will use. Employers should also inform employees of the individual, vendor, or entity that created the system and that will run, manage and interpret its results. A copy of this notice should also be provided to the labor agency within 10 days of distribution to workers, and notices should be updated following any significant updates or changes to the system.
Employers that develop, procure, use or implement an ADS or worker information system are required to complete an algorithmic impact assessment or data protection impact assessment, respectively Impact assessments must occur before the use of the system, or retroactively for systems in use at the time of the legislation coming into effect, and should be conducted by an independent assessor with the relevant experience and understanding of the system.
The aim of an impact assessment is to evaluate the potential risks posed by a system. These include discrimination against protected classes, violations of legal rights, direct or indirect physical or mental harms for algorithmic systems, and privacy harms for worker information and algorithmic systems. Assessors should also identify whether a system could have a chilling effect on workers exercising their legal rights or a negative economic or material impact on workers. Impact assessors must also assess whether a system has the potential to infringe on the dignity and autonomy of workers and errors (false positives and negatives).
When conducting either type of impact assessment, the assessor or auditor is required to consult workers who are potentially affected by the ADS or worker information system under investigation. This can include the identification of the risks to be evaluated and mitigation strategies to mitigate these risks. The assessor should also make the preliminary assessment available to workers for anonymous review and comment, and employers are prohibited from retaliating against workers who participate in this.
Employers of workers that operate from a workplace in California who collect data about their workers, use electronic monitoring, or use automated employment decision tools to make employment-related decisions about workers. Vendors who act on behalf of employers also share liability and must comply.
Written by Airlie Hilliard, Senior Researcher at Holistic AI & Ayesha Gulley, Public Policy Associate at Holistic AI.
Subscribe to our newsletter!
Join our mailing list to receive the latest news and updates.
Our automated AI Risk Management platform empowers your enterprise to confidently embrace AIGet Started