The Arlosoroff Forum, the Histadrut’s research and policy institute focused on the labor market and the economy and formulates proposals to improve workers' conditions and rights, is calling on the government to establish a round table within the National Artificial Intelligence Directorate, with representation from the state, the Histadrut, and employers.
According to the Forum, this round table will work to map the risks and opportunities arising from the integration of artificial intelligence in the labor market, examine regulatory and contractual solutions developed in other countries, formulate skills forecasts, develop pilots and practical tools for responsible deployment in workplaces, and track the technology's impacts on employment, wages, productivity, and inequality.
The Demand: Involve Workers and Employers in Policy Shaping
In an appeal sent about a month ago by the Arlosoroff Forum to the Chair of the Subcommittee on Artificial Intelligence and Advanced Technology, MK Orit Farkash-Hacohen, the Forum warned that advancing the national AI plan without involving workers and employers risks failing to realize the economic and social potential of artificial intelligence, and could even widen skills gaps and deepen inequality among workers.
"In such a scenario, workers, employers, and the entire economy will bear the costs of the transition, while the fruits of innovation will concentrate in the hands of a narrow slice of companies and workers," wrote Alon Marcus, Policy Advocate at the Forum, and researcher Shay Weinblum.
In the letter, Alon Marcus and Shai Weinblum welcomed Government Decision 4255 to promote the national AI plan, but emphasized that the plan lacks sufficient preparation for adapting and training Israel's human capital to the changing reality, nor does it include adequate regulation to ensure that AI leads to inclusive growth.
"Advanced infrastructure and technological leadership do not, in themselves, guarantee that the implementation of artificial intelligence will lead to higher-quality employment and narrowed gaps," they noted. "These outcomes are not predetermined; they depend on how the state designs and prepares for the integration process in the labor market, together with workers and employers. Among other things, this includes how tools will be developed for changes in work content, required skills, and management methods: who gets hired and who is filtered out; how tasks and shifts are allocated; how performance is measured; which employees are promoted; who is subject to surveillance; and under what circumstances decisions are made regarding position reductions or termination of employment."
The two emphasized that under the current proposal, there is no permanent framework for involving workers and employers in policy-making. "Representative labor organizations possess vital knowledge regarding work processes, skills, safety, and the implications of organizational change. Research by the OECD and the ILO shows that social dialogue facilitates a more effective, equitable, and accepted implementation of artificial intelligence."
Under the proposal, a round table of the state, labor organizations, and employers would be established, incorporating experts from various disciplines. The role of the round table would be to map the risks and opportunities of AI deployment in the labor market, examine regulatory and contractual frameworks developed abroad, formulate skill projections, develop practical pilots and tools for responsible adoption in workplaces, and monitor the technology's effects on employment, wages, productivity, and inequality. The round table's work would culminate in an annual report submitted to the government and the Knesset.
Training, Skills, and Protecting Workers' Rights
Alongside this, they highlighted that the government decision currently lacks a mechanism for training and upskilling existing workers in the economy. The challenge is not limited to training tech-sector workers, but extends to adapting the skills of existing employees across all sectors, with an emphasis on low-wage workers. Research by the Arlosoroff Forum on Lifelong Learning indicates that labor organizations worldwide actively participate in managing national training systems and negotiate over training opportunities.
The research also points to the ability of labor organizations to boost worker participation in these training programs and adapt them to their needs. Consequently, the Arlosoroff Forum proposes revitalizing the Employers and Human Capital Development Directorate—a joint initiative of the state, employers, and the Histadrut—to fund training and career transitions in sectors exposed to rapid technological change, prioritizing workers and demographics at higher risk. The mechanism would focus on on-the-job training, skills adaptation, assistance to businesses, and professional re-skilling.
Additionally, they emphasized that while AI systems can improve job matching, safety, and productivity, they can also give rise to discrimination, privacy violations, and decisions that are inexplicable or unappealable. The government decision fails to provide a sufficient answer to these risks, necessitating clear rules and regulation to guarantee fair and appropriate use of AI in workplace decision-making.
Therefore, it is proposed to mandate the Ministry of Labor, the Ministry of Justice, the Equal Employment Opportunities Commission, and the Privacy Protection Authority, in coordination with representative labor organizations and employers, to formulate a framework for responsible AI use in the workplace. This framework would include, among other things:
- An obligation to inform employees and candidates when AI systems are used in decisions affecting them.
- The right to an explanation, human review, and an appeal process.
- A strict prohibition on direct and indirect discrimination.
- Protection of privacy and personal data.
- Mandatory impact assessments prior to deploying significant AI systems.
- An evaluation of liability for both the employer and the technology vendor.
Distribution of Productivity Gains and Public Service Preparedness
Artificial intelligence has the potential to boost productivity and output per worker, but there is no guarantee that productivity gains will translate into higher wages, reduced working hours, or improved working conditions. "Without institutions for collective bargaining, sectoral regulation, and worker partnership, a significant portion of economic value risks concentrating among corporations, capital owners, and high-skilled workers." OECD research in surveyed countries found that positive impacts on employment and wages tend to be stronger among high-wage, high-skilled workers, while lower-skilled workers benefit far less.
Regarding the public sector, the Arlosoroff Forum warned against using artificial intelligence as a substitute for investment in public services.
Furthermore, forum research found that the wage gap between Israel's public and private sectors stood at approximately 27% in 2023—among the highest of the countries surveyed. "Against the backdrop of this extraordinary wage gap between the public and private sectors and the difficulty in attracting quality personnel, deploying AI without investing in workers risks deepening the weakness of the public service. The state acts both as a regulator and as the largest employer in the economy; therefore, it must serve as an example for leading this critical issue. It is proposed to formulate a binding government procedure stipulating that any AI system deployment in public service affecting state employees must require regulation."
Finally, the Forum pointed out that Israel currently lacks a structured database to systematically monitor the effects of artificial intelligence on the labor market. "There is a lack of data on the sectors and occupations where AI systems are being integrated, the changes they create in task content, which workers receive training versus those left behind, and their impact on wages, working hours, job quality, and employee turnover. Additionally, there is insufficient tracking of potential gaps between the center and the periphery or among different population groups, as well as impacts on privacy, discrimination, safety, and burnout. Without such an information infrastructure, the government will struggle to identify risks in real time and proactively manage the transition, rather than merely reacting to changes after they have already occurred."