IBM Japan Settles Dispute Over AI-Based Personnel Evaluations
A landmark agreement has been reached between the Japan Metal, Manufacturing, Information and Telecommunication Workers' Union (JMITU) and IBM Japan Ltd., resolving a labor dispute over the use of artificial intelligence (AI) in personnel evaluations.
The agreement, approved by the Tokyo metropolitan government's Labor Relations Commission on August 1st, requires IBM Japan to disclose to the union the data used by its AI system to evaluate employee performance and work attitude. This information will be used to assess the fairness and transparency of the AI-based evaluations.
The agreement comes after the union filed a petition for a relief order in April 2020, citing concerns about the AI system's potential for invasion of privacy, discrimination, and automation bias. The union argued that without full disclosure of the AI's decision-making process, there could be no meaningful negotiations regarding its use.
IBM Japan initially resisted disclosing the AI-related information, claiming that final evaluation decisions were made by human managers and that the AI served only as a reference. However, the company has now agreed to provide the union with detailed information about the AI system, including the 40 factors it considers in salary assessments and the relevance of these factors to the evaluation process.
The union hailed the agreement as a model for protecting workers' rights and working conditions in the age of AI. They believe the agreement will encourage other workplaces using similar AI systems to adopt similar transparency measures.
Yosuke Minaguchi, a lawyer representing the union, emphasized the significance of the agreement, stating that it is the first of its kind in Japan and could have a significant impact on other companies using AI for personnel evaluations.
The agreement comes amidst growing international concern about the ethical and legal implications of AI. The European Union recently passed the world's first Artificial Intelligence Act, which classifies AI-based personnel evaluation as a high-risk activity requiring strict management. Japan is also considering legal regulations for AI use.
This agreement between JMITU and IBM Japan sets a precedent for ensuring transparency and accountability in the use of AI for personnel evaluations, paving the way for a more ethical and responsible approach to AI implementation in the workplace.