Ai in HR, what every organisation must get right before deployment

DOWNLOAD REPORT

Ai in HR – what every organisation must get right before deployment

Authored by Mandeep Singh, Partner – HR, AI & Data Science

As the Ai wave is covering various aspects of our work life, and business process after business process get covered, it is matter of time before human resources function also has Ai embedded into it processes. So, if you are planning to introduce or have already introduced Ai in HR processes, this article is a must read.

This paper covers

• Implementation plan
• Legal aspects to keep in mind
• Points to remember
• HR Ai Reliability Framework

First and foremost, we need to be clear about what you will use Ai for, that is

For Recruitment

It means, you could possibly use Ai to assist in resume screening, candidate interviewing, analyzing content of interviews, for making shortlisting decisions, for conducting various kinds of test, irrespective whether they are technical or behavioral or for making the final hiring decision based on all the data and information provided, even if that decision is just saying a yes or a no, and the final call is taken by the HR practitioner.

For Talent Management and Capability

It means you plan to use the Ai for evaluating the comments of evaluators, taking out key themes from employee self-evaluation forms or reviewing the evaluation forms and inferring meaning out of them. It means, it could be used to analyze training needs, to understand skills, for gap analysis or identifying potential in an employee or to decide performance ratings, even to the level of thinking who may be meeting eligibility criteria and who may not be meeting the criteria, and in some cases even recommend people for promotions or state who is not eligible for promotion.

For Payroll, compensation and employee document analysis

It could also be used for doing payroll calculations, for fixing bands, compensation ranges or for fitting people into automatically created compensation structures which are tax or benefit friendly. We could also use if for collecting employee documents, arranging them into HR files, creating employment records, or even just reading these documents and entering them into tables of an HRMS

and maybe, for Sentiment and other analytics

This is, are you going to be using it for predictive analysis, like, will the person leave in a particular time, will they be good, bad or average in a particular job, will two people get along in the team, whether to assign them a particular job or not, gauging whether the person is in a good mood, bad mood or behavior analysis, or monitoring work habits and behavior by analyzing key strokes, emails, attendance pattern, or even their face recognition

As you prepare to get your thoughts together on what you will use Ai for, it is important to understand the evolving landscape and how the journey of Ai should be introduced in organisations.

As the rules of engagement are still evolving, it puts the onus of responsible use entirely on the user, so let us talk about a few important developments that can help us implement Ai responsibly in our organisations.

Let’s get the implementation plan

Who champions it and making of the Ai policy

First things first, Ai usage and implementation is not so much a technical decision as it is an ethical and moral decision. So, you need a team comprising of a technical expert, a HR Tech expert along with your organisations ethics and integrity committee (which would ideally also be having the legal team in it).

The first and primary task of the Ai committee, or whatever name you call it, is to draft the Ai policy for your organisation, with the scope and usage of Ai clearly defined. This important step discusses things like what the Ai systems can do, what checks and balances will be put in place, how data can be collected and put into Ai system, how the organisation will inform employees about Ai adoption and most importantly, how it impacts their career and experience at the organisation.

It doesn’t stop there, it further plans out a system of how there will be continuous monitoring and implementation as the world of Ai evolves and what use cases will be brought into the experiment, do not try and completely embed into Ai domain.

Primarily this steering group is creating a full business plan on how Ai will be embedded into the organisation, and like any business plan this is not an individual’s decision, but a senior management decision.

The question that gets asked many times is why should the ethics and integrity committee, or HR Tech expert, or for that matter, others outside of the HR function get involved, specially as most other departments may just discuss within their department and take a call of introducing Ai in their processes.

The answer is simple, this is a people decision, which is supported by the internal HR function as advisors, they don’t own people processes like any other business process, and we are dealing with humans, which is much larger than a business outcome, so this is an area where ethics committee definitely gets involved.

Then comes fairness

When we talk fairness, we basically mean that on a continuous basis, we need to keep running tests and simulations to be aware what outputs are coming using the Ai system. What are the deviations, what is the change and what is the evolving acceptable limit. A continuous monitoring is necessary to keep giving a report every few months to be aware of what Ai is doing and how it is deciding on the identified use cases.

For example, if the organisation uses Ai for resume screening, or creating a training needs analysis, then constant evaluation and testing needs to happen as parallel process on an ongoing basis for years and not on a sampling basis.

Then comes communication

Communicating is a necessary part, not only to employees, but also to the outside world. The use case, its impact, the ability to question and get answers from the Ai process and its outcomes need to be informed not only to employees, but also to any individual who is exploring career or has their data being processed through the Ai system.

More importantly this communication is supposed to cover how their data is being used, collected and retained. As in, will the company be using the data of the users to train the Ai system, to train the model (or create a model) and can a user have a right for erasure, and if that right is still with the employee or the user, then how will the organisation manage such requests.

Why a focused approach is necessary, the legal aspects

The laws, guidelines and appropriate ways of working using Ai in HR are still developing, which means we should progress using those processes which will not have a negative impact on the organisations reputation or be forced to make a system built with considerable amount of resources being made redundant as a result of some legislation. Therefore, we rely on certain guidelines which have evolved and which seem to be a reasonable demand from any HR Ai system.

Let us try and discuss which are the guidelines which have made some kind of impact

The first is the Ai guidelines and principles emerging from the International Labour Organization (ILO)

Since ILO primarily works to protect workers rights, its focus is on ensuring that Ai enhances productivity without undermining fairness, dignity or employment rights. Since labour laws in many countries have been influenced by ILO, there is a reasonable possibility that future employment related Ai regulations will draw heavily from these principles.

The second is the EU Ai Act

Like many European regulations, it has a precautionary approach, placing importance on transparency, accountability, fairness, and risk management. It is one of the most comprehensive and detailed Ai governance frameworks available in the world today. Unlike other guidelines, this is a law within the European Union. So, organisations that engage with the European Union will need to align their Ai practices with these requirements. Even organisations outside European Union are referring to this Act, as it provides a blueprint for responsible Ai governance.

The third are the OECD Ai Principles

These have become one of the most influential global reference points for Ai governance. Although, unlike the EU Ai Act they are not legally binding, they have been adopted or endorsed by many governments and international bodies. The OECD principles focus on trustworthy Ai, encouraging organisations to ensure transparency, accountability, fairness, robustness, and human oversight in the design and deployment of AI systems.

The fourth is UNESCO’s Recommendation on the Ethics of Artificial Intelligence

Unlike other regulations UNESCO approaches Ai from the perspective of human dignity, human rights and the long term impact of technology on society. While these principles are not directly enforceable as laws, they represent the ethical foundations upon which many future Ai regulations are likely to be built as these guidelines provide an moral and ethical compass.

Points to remember

We recommend the undermentioned that every organisation should follow when implementing a Ai system in HR

• Ensure that ownership and accountability are clearly allocated within the organisation for every decision taken by the Ai system, and the ownership of the decision is always with the person. That means Ai may support HR, but the employee/manager remains accountable for screening, hiring, shortlisting, promotion, appraisal, compensation, disciplinary and termination decisions.
• It may be the case that sometimes humans would choose to rely on AI systems for reasons of efficacy, but the decision to cede control in limited contexts remains that of humans, as humans can resort to AI systems in decision-making and acting, but an AI system can never replace ultimate human responsibility and accountability, UNESCO Recommendation on the Ethics of Artificial Intelligence (2021)
• Member States should ensure that it is always possible to attribute ethical and legal responsibility for any stage of the life cycle of AI systems, as well as in cases of remedy related to AI systems, to physical persons or to existing legal entities, UNESCO Recommendation on the Ethics of Artificial Intelligence (2021)
• To this end, AI actors should implement mechanisms and safeguards, such as capacity for human agency and oversight, including to address risks arising from uses outside of intended purpose, intentional misuse, or unintentional misuse in a manner appropriate to the context and consistent with the state of the art, OECD Recommendation of the Council on Artificial Intelligence (2024 version)
• Deployers shall assign human oversight to natural persons who have the necessary competence, training and authority, as well as the necessary support, EU AI Act / Regulation (EU) 2024/1689
• Test for bias before and after deployment, that is on a continuous basis. On a continuous basis keep testing, evaluating and observe how the Ai system is working and where required modify, or stop it
• AI actors should make all reasonable efforts to minimize and avoid reinforcing or perpetuating discriminatory or biased applications and outcomes throughout the life cycle of the AI system to ensure fairness of such systems. Effective remedy should be available against discrimination and biased algorithmic determination, UNESCO Recommendation on the Ethics of Artificial Intelligence (2021)
• AI actors should respect the rule of law, human rights, democratic and human-centred values throughout the AI system lifecycle. These include non-discrimination and equality, freedom, dignity, autonomy of individuals, privacy and data protection, diversity, fairness, social justice, and internationally recognised labour rights, OECD Recommendation of the Council on Artificial Intelligence (2024 version)
• Deployers shall monitor the operation of the high-risk AI system on the basis of the instructions for use and, where relevant, inform providers in accordance with Article 72. Where deployers have reason to consider that the use of the high-risk AI system in accordance with the instructions may result in that AI system presenting a risk within the meaning of Article 79 (1), they shall, without undue delay, inform the provider or distributor and the relevant market surveillance authority, and shall suspend the use of that system, EU AI Act / Regulation (EU) 2024/1689
• Deployers of high-risk AI systems shall keep the logs automatically generated by that high-risk AI system to the extent such logs are under their control, for a period appropriate to the intended purpose of the high-risk AI system, of at least six months, unless provided otherwise in applicable Union or national law, in particular in Union law on the protection of personal data, EU AI Act / Regulation (EU) 2024/1689
• Disclose what will Ai do, and how it will impact the employee and career aspirants when engaging with your organisation.
• People should be fully informed when a decision is informed by or is made on the basis of AI algorithms, including when it affects their safety or human rights, and in those circumstances should have the opportunity to request explanatory information from the relevant AI actor or public sector institutions, UNESCO Recommendation on the Ethics of Artificial Intelligence (2021)
• AI Actors should commit to transparency and responsible disclosure regarding AI systems. To this end, they should provide meaningful information, appropriate to the context, and consistent with the state of art: … ii. to make stakeholders aware of their interactions with AI systems, including in the workplace, OECD Recommendation of the Council on Artificial Intelligence (2024 version)
• Before putting into service or using a high-risk AI system at the workplace, deployers who are employers shall inform workers’ representatives and the affected workers that they will be subject to the use of the high-risk AI system. This information shall be provided, where applicable, in accordance with the rules and procedures laid down in Union and national law and practice on information of workers and their representatives, EU AI Act / Regulation (EU) 2024/1689
• Have a person who can answer questions with reasons when people are affected by Ai supported decisions for anything related to career, the person should be able to provide an explanation, and not simply say, this is as per the Ai
• In addition, individuals should be able to access the reasons for a decision affecting their rights and freedoms, and have the option of making submissions to a designated staff member of the private sector company or public sector institution able to review and correct the decision, UNESCO Recommendation on the Ethics of Artificial Intelligence (2021)
• AI Actors should commit to transparency and responsible disclosure regarding AI systems. To this end, they should provide meaningful information, appropriate to the context, and consistent with the state of art: … iii. where feasible and useful, to provide plain and easy-to-understand information on the sources of data/input, factors, processes and/or logic that led to the prediction, content, recommendation or decision, to enable those affected by an AI system to understand the output, and, iv. to provide information that enable those adversely affected by an AI system to challenge its output, OECD Recommendation of the Council on Artificial Intelligence (2024 version)
• AI actors should ensure traceability, including in relation to datasets, processes and decisions made during the AI system lifecycle, to enable analysis of the AI system’s outputs and responses to inquiry, appropriate to the context and consistent with the state of the art. AI actors, should, based on their roles, the context, and their ability to act, apply a systematic risk management approach to each phase of the AI system lifecycle on an ongoing basis and adopt responsible business conduct to address risks related to AI systems, including, as appropriate, via co-operation between different AI actors, suppliers of AI knowledge and AI resources, AI system users, and other stakeholders. Risks include those related to harmful bias, human rights including safety, security, and privacy, as well as labour and intellectual property rights, OECD Recommendation of the Council on Artificial Intelligence (2024 version)
• Avoid tools that infer stress, engagement, mood, loyalty or emotional state of employees from face, voice, keystrokes or other biometric/behavioural signals
• The following AI practices shall be prohibited: … (f) the placing on the market, the putting into service for this specific purpose, or the use of AI systems to infer emotions of a natural person in the areas of workplace and education institutions, except where the use of the AI system is intended to be put in place or into the market for medical or safety reasons, EU AI Act / Regulation (EU) 2024/1689
• Have a system which checks and classifies what data goes into the Ai system. Do not upload irrelevant, incomplete or biased HR data into an AI decision workflow.
• Without prejudice to paragraphs 1 and 2, to the extent the deployer exercises control over the input data, that deployer shall ensure that input data is relevant and sufficiently representative in view of the intended purpose of the high-risk AI system, EU AI Act / Regulation (EU) 2024/1689
• Receive a proper instruction manual on how to use the system in plain simple language
• High-risk AI systems shall be accompanied by instructions for use in an appropriate digital format or otherwise that include concise, complete, correct and clear information that is relevant, accessible and comprehensible to deployers, EU AI Act / Regulation (EU) 2024/1689
• Be very careful that the data of the employee being used in the Ai complies to the DPDP Acts, employment purpose, and in other places, do take consent
• A person may process the personal data of a Data Principal only in accordance with the provisions of this Act and for a lawful purpose,— (a) for which the Data Principal has given her consent; or (b) for certain legitimate uses. For the purposes of this section, the expression “lawful purpose” means any purpose which is not expressly forbidden by law, Digital Personal Data Protection Act, 2023
• Maintain data accuracy, if that is not the case, as HR you are not doing your job
• Where personal data processed by a Data Fiduciary is likely to be— (a) used to make a decision that affects the Data Principal; or (b) disclosed to another Data Fiduciary, the Data Fiduciary processing such personal data shall ensure its completeness, accuracy and consistency, Digital Personal Data Protection Act, 2023

In the end, I would just ask you to answer two questions

• Can you clearly explain where Ai is being used in HR, who is accountable for it, and how employees are affected by it?
• What are the ratings your Ai systems have on the HR Ai reliability framework

If your questions are answered positively, it is good, if not, you need to get your act together, remember this is not a HRMS, this is different.

DOWNLOAD REPORT