Essential legal considerations for AI implementation & employee rights

Essential legal considerations for AI Implementation and employee rights
AI systems are trained to make decisions based on patterns in data, and they can pick up inherent biases present in the algorithms that train them.

The times we live in are ever-changing, and the exponential rise in the use of artificial intelligence (AI) is proof of this. The increasing utilization of AI spans across various business verticals, from recruitment and performance evaluation to marketing, customer services, quality assurance, factory automation, and more.

As organizations heavily rely on AI, valid concerns regarding job loss, bias, and AI transparency have been raised. This article delves into the intricate relationship between AI implementation and employee rights from an Indian legal perspective.

Notice requirement for implementing AI

Organisations looking to include AI in their employment processes must also consider the implications of the implementation of such AI.

According to the Industrial Disputes Act of 1947, the law that governs changes in conditions of service, layoffs, retrenchment, etc., for “workman” category employees – usually referring to non-managerial and non-supervisory staff – employers must adhere to a specific procedure before implementing any rationalization, standardization, or improvement of the plant or technique that might result in the retrenchment of workmen.

This law requires that if such change is contemplated by the employer, then the “workmen” be intimated of introduction of such change at least 21 days in advance.

While Indian courts are yet to address the issues concerning the applicability of the said provision of law to the specific instances of the introduction of AI and the ever-changing business requirements, AI being a form of technology, the extension of the said legal requirement for cases seeking implementation of AI that may lead to a reduction in workforce, cannot be disregarded.

In a case where the introduction of a certain machine in the operations of the establishment led to the termination of employment of a workman, the Apex Court said that 21 days’ notice ought to have been given to the workman before introducing the machine, as termination of employment was a direct consequence of the introduction of the full-fledged machine[1]. 

In another case, the High Court of Delhi[2] held that, “…obviously a notice of change was a must before introducing the change, otherwise it would be an illegal change… It must be held that without anything more such an illegal change would be wholly ineffective…”

In this case, the retrenchment of employment of the workman was held illegal for non-compliance with 21 days’ notice requirement before computerising the process that was carried out manually.

The upcoming Industrial Relations Code, 2020, contains similar legal requirements, and the concerns around the procedure to be followed before introducing AI with the potential to replace manpower will continue until put to rest by the legislature or Apex Court.

Other legal aspects for consideration

AI systems are trained to make decisions based on patterns in data, and they can pick up inherent biases present in the algorithms that train them. Murkier waters lie ahead if the AI is trained with biased data or algorithm.

Candidates from marginalised communities will face greater challenges if such unfair methods are adopted by organisations. This bias may also have a significant impact on the evaluation of employee’s performance.

Biased systems may provide subjective assessment, and employees may face unfair judgement, disparities in feedback, and biased decisions in case of promotions or the advancement of their career. The impact of this can run very deep, as discriminatory practises may lead to hostile work environments with employees being demotivated, anxious, and even feeling harassed at their workplace.

Hence, to ensure that AI is used in an effective and non-discriminatory manner, organisations need to establish clear guidelines about the use of AI and create a system where employees’ concerns about bias can be addressed and answered. Organisations also need to keep a close watch on their AI systems to identify and rectify biases that may arise over time.

In India, apart from the protections against discrimination afforded to citizens in the Constitution, there are a number of legislations that prevent discrimination of any kind, such as – the Rights of Persons with Disabilities Act, 2016 provides guidelines to certain establishments relating to the employment of persons with disabilities and prohibits such establishments from discriminating against such persons, the Transgender Persons (Protection of Rights) Act, 2019 also prohibits discrimination by any person or establishment against transgender persons.

Organisations must also adhere to the guidelines in the Equal Renumeration Act, 1976, and The Code on Wages, 2019 (yet to be brought in force), which do not allow for a discrepancy in renumerations for men and women doing the same job.

Employers and organisations are obligated to provide their employees with a fair and non-discriminatory work environment.

Another key area of concern is the privacy of personal data collected by organisations, and AI should be compliant with the data protection rules of the land.

Need for a transparent system

It is important that an organisation relies on AI, the process of which is easily explainable. Black box AI whose inputs and operations aren’t visible to the user or provides decisions without providing any explanations as to how they were reached, result in a complexity that may be difficult to understand.

This lack of transparency may lead to there being a significant distrust of AI by employees and resultant differences. Transparency regarding the use of AI allows employees to understand the metrics that are used to make a decision by an organisation.

With the use of AI drastically expanding, organisations should look to use algorithms that they can understand, and they should keep their employees in light of the algorithms that are being used.

Looking forward

While the benefits of AI are obvious, the questions, concerns, and consequences of its implementation require a thorough evaluation to ensure organisational preparedness to address any valid questions from employees.

Nevertheless, the fear and concerns about the introduction of AI in the system are genuine, and the legal requirements that may be triggered around such an introduction may be far-reaching. A detailed analysis of AI, its process, and its consequences is a must before implementing it.

Co-Authored by Shrivar Bajoria


Please enter your comment!
Please enter your name here