Governments should consider explicitly regulating or prohibiting electronic monitoring of an employee’s activities without prior notice to all employees who may be affected.

Governments should evaluate the extent to which psychometric testing—the use of questionnaires aimed at psychologically profiling job candidates—provides accurate insights about whether someone is qualified for a job. Decision-makers should consider regulating or limiting employers' use of these tools to make employment decisions based on their potential to create discriminatory outcomes.

Governments should rigorously evaluate, and consider regulating, employers' use of predictive algorithmic software to make decisions about job recruitment, hiring, compensation, and employee evaluation. In light of the growing prevalence of these tools, which create the potential for unfair profiling and discrimination, decision-makers should have accurate information about what inputs and functionalities produce these outcomes.

Governments should require employers to facilitate workers’ communication by providing workers (and/or worker organizations) with a way to contact their coworkers. This obligation could run either to everyone in the organization, a default that many companies currently use (e.g., because their internal email address books include everyone in the organization); to all of the workers in an organization who are covered by labor law; or to subsets of workers who do similar jobs. This requirement could be as straightforward as requiring employers to provide workers with a list of company email addresses or other contact information, but a communications-facilitation requirement could be more effective if workers had a private forum for online communication.

Law- and policy makers should consider new regulations that interpret anti-discrimination laws in light of predictive hiring tools. Predictive or algorithmic hiring technology, increasingly in use by major employers across the world, are subject to human bias and can affect equity throughout the entire hiring process. Current laws and enforcement frameworks must be updated to hold vendors of such technology accountable and ensure that workers and job applicants continue to be treated with fairness in recruiting, interviewing, and selection.

While some automated technologies raise productivity and deliver economic benefits that can help offset the impacts of displacement, others (labeled by some researchers as “so-so technology") may deliver limited benefit while adversely impacting workers. Governments and employers need more investigation and a more nuanced understanding of the merits of specific types of automation and technology, such as self-checkout. Defining “good” and “bad”—or, perhaps, “worse”—automation can be useful in regulating change, determining policy priorities, and deploying employer or worker incentives.

‍

Workers and labor unions should be partners in the process of implementing electronic health records, and policymakers can promote or require such participation.  One approach could be to model work councils in some European countries, where worker representatives are elected and the committees are given certain powers over a range of decisions related to technology. Alternatively, workers and unions could be granted greater influence over existing labor-management safety and health committees in some states in the U.S. Research documents superior outcomes in the implementation of electronic health records when workforce planning was integrated in the policy, operations, and design process.

Governments should pass regulations to mitigate the potentially harmful effects of new data collection and surveillance tools on job quality and human dignity. In addition to a general default legislative framework, unions should also push for tailored regulation through collective bargaining, since optimal rules for data collection and use may vary considerably among workplaces.

Firms should focus on the creation of better data environments to maximize the use of machine learning for public problem-solving. Efforts could focus on improving the systems and protocols by which data is defined, gathered, accessed and manipulated. This includes government initiatives for open public data, industry-government collaboration on data and code verification, or audits and policy frameworks (or agreements) to make strategic data available to specific users – with specified safeguards – in order to enable AI applications for societal and environmental benefits.

Firms should develop industry-wide and industry-regulator teamwork to aid in AI standard-setting (for example, through consensus protocols and smart contracts that include efficiency principles, or which require common agreement and governance).

Firms should establish board-level AI advisory units to ensure that their boards understand the implications of AI technology, including safety, ethics, and governance considerations.

Universities should offer more modules and courses focusing on the ethics in technological innovation. A systematic educational approach could help new tech leaders distinguish between technologies that expand the capability of workers and those that merely extract from them without increasing their potential.

Investors should create and abide by a Charter for Ethical Technology Investments so that there exists a widespread standard for investing in ethical innovations. Investors are increasingly active in shaping the behavior of tech firms, and their influence can successfully pressure large companies to take into account the potential risks of technology on workers.

Tech companies should publicly disclose the makeup of their workforce, including demographic and educational data. Improving diversity within tech firms would benefit the public—since a tech workforce that mirrors society is more likely to create appropriate products for the whole of society—and improve company performance.

Tech companies should give hiring preference to applicants who have taken ethical modules.

Technology firms need to be more cognizant and deliberate with their recruitment practices. Ways to increase recruitment from marginalized groups include blind testing of applicants and partnerships with advocacy organizations. Improving the diversity of tech employees will bring dividends to the public, because a workforce that mirrors representation in society will create technology that all people can use.