As organisations strive to enhance their efficiency and decision-making processes, the use of technology, including Artificial Intelligence (AI), is becoming increasingly commonplace.

In almost every industry, AI has revolutionized the way businesses operate, and Human Resource functions are no exception. By streamlining HR processes in a cost-effective manner, AI has the potential to provide numerous benefits to organisations. However, before implementing AI in their organisations, companies must take into account certain legal considerations.

These considerations involve ensuring that the AI systems used in HR functions comply with applicable laws, such as those related to non-discrimination, data privacy, and cybersecurity. While the incorporation of AI in HR functions can bring about increased efficiency, accuracy, and cost-effectiveness, it also brings with it the responsibilities of a data processor, which places companies in a realm of uncertain and evolving legalities.

Legal aspects to be considered before adopting AI in HR

AI in the HR domain will speed up and rationalise the HR process in a cost-effective manner. However, there may be larger implications for businesses if AI breaches any law of the land. 

Fundamental freedom of an Indian citizen to practice any profession, or to carry on any occupation, trade, or business is envisaged under Indian Constitution. Indian law also specifically prohibits discrimination based on gender in employment, and many job roles are still dominated by men. Further, separate legislations for the protection of rights of the trans-persons and HIV-positive persons are enacted in India to protect their rights and promote their welfare, including protection against discrimination and unfair treatment at employment. Indian law also prohibits discrimination and unfair treatment of specially-abled persons in certain employments. Hence, the AI adopted in the selection and recruitment process should be trained to be non-discriminatory as required under applicable laws. This requires avoiding bias in AI algorithms and ensuring that the data used by AI systems are representative of the entire eligible population. The employers must also ensure that the AI system's decision-making processes are transparent and explainable so that the individuals affected by the decision made by the employer using AI can understand the reasoning behind the decision of the company.

The AI systems should be designed to be accessible to everyone qualified for the purpose. In a recent case, the High Court of Bombay upheld the right of the trans-persons to apply for the job of police constable and directed the state government to facilitate the submission of applications online by transgender persons within a stipulated timeframe. Hence, the adoption of AI in HR processes will not dilute the employer’s obligations to ensure a non-discriminatory and fair work environment as required under applicable laws.

Further, the Information Technology Act, 2000 (the “IT Act”) provides for the protection of personal information and privacy of individuals, and HR professionals must ensure that the AI system used in HR functions complies with these provisions.

Under the IT Act, the entities must obtain the consent of employees before collecting, processing, or storing their personal data, including sensitive personal data such as passwords, medical records and history, biometric information, certain financial information, etc. The entities must also ensure that the AI system used in HR functions does not collect or process any personal data without the employee's explicit consent. Additionally, it is necessary that the AI system's data processing activities are in line with the principles of purpose limitation, data minimization, and storage limitation. 

Under the IT Act, the organisation should use reasonable security practices and procedures to protect personal data from unauthorized access, use, disclosure, modification, or destruction. This includes implementing appropriate technical and organizational measures to safeguard personal data from cybersecurity threats such as hacking, malware, and phishing attacks. The responsibility for the safety of such data would be on the entities, whether they are using AI in their HR functions or not. 

Another important consideration is ensuring that sensitive personal data is not disclosed to unauthorized persons. HR professionals must also ensure that the AI system's access controls are robust and that the system's logs are regularly monitored to detect and prevent any unauthorized access or use of sensitive personal data or information.

Apart from the above, the complex labour law regime in India requires employers to be more cautious in employee management and there is no straightjacket formula for the resolution of employee disputes. The laws applicable vary basis the size of the organisation, location, nature of work of the employee, etc.; and the employee dispute management depends on the facts and circumstances of the case at hand. Needless to say, communication with employees plays a major role in employee dispute management. The High Court of Himachal Pradesh stated that the Courts should not behave like an artificial intelligence machine but should adopt a justice-oriented approach. Hence, the AI which cannot identify non-technical issues such as behavioural aspects or the emotional/psychological impact of an employer’s decision will require human interference. Any AI system adopted for the redressal of employee grievances cannot just be unique, but it must be dynamic and progressive.

While the implementation of AI in HR functions can bring numerous benefits to organisations, including increased efficiency, accuracy, and cost-effectiveness, it also puts upon the employer the myriad responsibilities of a data processor that bring it into a realm of uncertain and ambiguous legalities that are currently still evolving. Even the replacement of the IT Act is on the cards, having come into force at a time when the internet was in its nascency. Hence, it is essential to identify the processes for incorporating the AI with specific exceptions for a human interface on a need basis, subject to applicable laws.