Governments around the world are increasingly recognising the need for novel legislation to ensure the responsible deployment of AI in the workplace.
To benchmark UK activity, we are mapping international legislation that recognises the novel impacts of AI at the workplace. We have also outlined related principles of IFOW's Good Work Charter for each piece of legislation. Our aim is to support research and policymaking by creating a comprehensive online resource.
Legislation currently mapped includes:
This policy tracker has been compiled by Stephanie Sheir, IFOW’s researcher specialising in AI governance, and will be periodically updated. Please get in touch via email@example.com if you think there is anything missing.
The Act implements transparency requirements for and a prohibition on using protected characteristics in algorithmic decision making that determines individual access to important life opportunities.
Washington DC, USA
The Act focuses specifically on algorithmic decision making with 'adverse action', defined as a denial, cancellation, or other adverse change or assessment regarding an individual’s eligibility for, opportunity to access, or terms of access to important life opportunities. The world of work would certainly be implicated here, and would be subject to novel requirements for notice and disclosure of the use of personal information in algorithmic decisions and auditing for discriminatory processing.
Equality, Access, Fair Conditions, Dignity, Wellbeing
The bill requires employers to conduct bias audits and disclose in advance the existence and assessment features of automated employment decision tools.
New York City, USA
Placed on agenda in 2020. Enacted in 2021. Taking effect in 2023.
The bill uses a precise but not overly narrow definition of automated employment tools and mandates sensible requirements surrounding non-discrimination and transparency disclosure to candidates.
Equality, Access, Fair Conditions
The Bill prohibits employers from using facial recognition in interviews unless written consent is provided by the candidate.
While notice and consent is insufficient on its own to safeguard applicants' rights, this is a useful first step in securing transparency and accountability in AI video interviews at the state level.
Dignity, Access, Autonomy, Participation
To ensure that automated decision-making systems used in the public sector are deployed in a manner that reduces risks to Canadians and federal institutions, and leads to more efficient, accurate, consistent and interpretable decisions made pursuant to Canadian law.
Passed, in third review as of April 2022.
The third review of the Directive proposes expanding the scope to include impacts on internal federal employees, effectively including the public sector workplace as an area of concern.
The Act regulates the use of AI to analyse video interviews of job applicants.
Likely the first of its kind, the Act implements accountability and transparency measures for AI video interviews, including disclosure of the use and functioning of AI. The Act requires the consent of job applicants, procedures for deletion of videos on request, restrictions on the sharing of videos and annual reporting of the race of hired and rejected applicants to the Department of Commerce and Economy Opportunity.
Access, Equality, Dignity
Described as the American GDPR, the ADPPA seeks to provide consumers with data protection rights such as data minimisation and the rights to consent and object, as part of a 'privacy by design' approach.
Introduced in the House of Representatives
The ADPPA mandates the use of Algorithmic Impact Assessments (AIAs) for large data holders that use an algorithm that may cause harm to an individual. These harms include those related to the making or facilitating advertising for, or determining access to employment, and those related to disparate impact on the basis of sensitive characteristics.
Section 207: Civil Rights and Algorithms
Access, Equality, Dignity
An act to amend Government and Labour Code.
Referred to the Commission on Privacy & Consumer Protection, California State Assembly
The Act places restrictions on the use of automated decision-making in the workplace, and mandates the disclosure of certain information by the employer to employees when these systems are deployed.
Access, Equality, Dignity
To direct the Federal Trade Commission to require impact assessments of automated decision systems and augmented critical decision processes, and for other purposes.
The Act is one of the most wide-ranging pieces of algorithmic accountability legislation, regulating automated decision-making that has any legal, material, or similarly significant effect on consumer lives. Although the Act is aimed at consumers and is thus enforced by the Federal Trade Commission, such general purpose legislation will certainly have impacts on the world of work and names employment as a critical area of concern.
A set of measures to improve the working conditions in platform work and to support the sustainable growth of digital labour platforms in the EU.
The Directive proposes novel wide-ranging rights for platform workers, with provisions on employment status, algorithmic management, human oversight of automated decisions and human review of significant decisions. Remedies for platform workers are also proposed, including rights to redress, protection from dismissal and communication channels amongst workers.
Access, Fair Conditions, Equality, Dignity, Support, Participation
The first law in the world seeking to govern AI by a major regulatory body.
The EU AI Act adopts a risk-based, horizontal regulatory approach to regulate AI, using a fixed, broad definition of AI and a product safety approach. The Act will have wide-ranging implications for all uses of AI on the single market, including work which is identified as a high risk area.
First-of-its-kind legislation in China regulating the targeting of recommendation algorithms.
The legislation establishes several novel rights for consumers, including the right to be informed about the use and operational mechanisms of algorithmic recommendation, the right to opt-out and the right to not to be subjected to differentiated treatment, which effectively outlaws 'unreasonable' dynamic pricing. Whether and how this applies to workers remains to be seen. The Bill requires service providers to cater to the rights and interests of elderly people and workers.
Access, Fair Conditions, Equality