By Ayan Paul
It is often said that “Data knows better.” In the real world, where most decisions are based on best practices or best guesses, this is especially true when it comes to engineering a company’s growth, be it through workforce transformation or hiring talent. However, there have been numerous and often heated debates about whether an algorithm’s decision-making abilities are better than a human’s. The differences between the two have been increasingly thinning out with the advent of robust AI algorithms and models that far outperform humans in several domains. Despite these successes, there has been skepticism about passing on the reigns to intelligent algorithms to make decisions that even humans need help to make, especially when they are non-quantifiable to a large extent. Quoting Freiherr von Eichendorff, a 19th-century poet and novelist:
“Quod licet Iovi, non licet bovi” – a Latin phrase that means: What is permissible for Jupiter (the God) is not permissible for cows.
The attitude has been similar in accepting data-driven AI methods in decision-making related to workforces. The threshold for acceptance of mistakes is far lower when a human (the algorithm creator) makes the decision than when algorithms are allowed to do so. This is primarily because we have come to expect perfection from anything that is automated, and rightfully so. One cannot allow for the absence of oversight over data-driven decision-making because there lies the distinct risk of harming individuals when human-centric decisions are made. Hence we see a reluctance and superlative caution in the allowance of full-fledged self-driving algorithms or the widespread use of AI in medicine and public health. According to the EU AI Act, “AI systems used in employment, notably for the recruitment and selection of persons … should also be classified as high-risk, since those systems may appreciably impact future career prospects and livelihoods of these persons.”
So what is our path forward for leveraging data in the decision-making process related to our workforce? Should we wait and watch, or have we already started integrating such methods in our companies? The answer is nuanced and layered. Data, analytics, and data-driven decision-making come at a cost. The biggest question is shall we get a return on investment if we incorporate these methods in our workflows related to our workforce?
The answer is that we should start building a future-proof company. This requires a definitive and well-planned strategy to augment HR tasks with data on its workforce and candidate pools that the company already has at its disposal. There are two main action areas:
- Workforce development
- Workforce augmentation through hiring
Each of these requires different solutions, but those that work together. The traditional methods of yearly appraisals and performance evaluation can be replaced by continual monitoring of the performance of each employee and their movement towards a specific goal. As demonic as it sounds to always watch individuals, this can provide employees with continual feedback to improve their skillset and reach their performance goals. It can also help reduce attrition due to a lack of performance or employee dissatisfaction.
However, keeping a human in the loop to oversee the feedback process and ensure the data-driven algorithms are performing according to expectations is best. This, in turn, requires a change in perspective of how HR and managers work together, a task that is necessary to make a company future-proof. Big-name HR brands provide such services, but much must be done to unbiased the process and make it fair and equitable for an employee, and not treat them less as just labour, which will lead to the holistic growth of the company instead of its operating like a factory where humans are expected to work like machines.
In Summary
In the realm of hiring, many of the mundane tasks performed by recruiters can be automated. Many products have come up recently to assist recruits, with many claiming to use AI and enable fair and equitable recruitment. The space many of these products need to address is tying the company workforce analytics to algorithmic recruitment decision-making. These two are often treated as disparate areas. There is a need for a combined algorithmic approach that helps strategize both action areas in an integrated and parametrically connected manner. The future lies in something other than separate teams making independent decisions hoping they will somehow be able to put a square peg in a round hole. Historically, this has been the largest source of dissipation of efficiency and growth in a company. It’s time to use data to bring the company together with a single workforce strategy augmented by data and AI. There is a lot of space for developing novel scoring methods that take into account more human aspects of the workforce and focus less on treating them as cogs in a machine.
The author is founder and chief scientific officer, KarmaV