28 Aug 2018
The notion of AI often conjures up images of automation without human involvement. But while it offers HR the potential for great leaps forward in organisational processes, human intervention and decision-making are still required.
The modern conception of AI has its antecedents in world mythology, through figures such as Galatea or the Golem. Later medieval and enlightenment thinkers would expand on the concept, directly or indirectly contributing to the development of the modern computer in the process. The idea of AI has also flourished in science fiction, with one of the most famous examples found in Mary Shelley’s Frankenstein.
But AI research as we understand it today emerged in 1956, at the Dartmouth Conference, an extended brainstorming session that helped lead to the modern understanding and definitions of AI. There are numerous subclassifications of definition to consider, but they consistently differentiate on task complexity.
“Arriving at a precise definition of AI is challenging,” says Darren Hnatiw, Chief Technology Officer at Frontier Software. “But the common factor in each is an intention to enable computers to behave in ways that mimic human behaviour and intelligence.”
AI has a huge number of potential applications within a business context, for example completing repetitive, low-value-add tasks. It offers tangible benefits, such as the capacity to assimilate and analyse large amounts of data, reducing human input time and the capacity for error. According to Deloitte’s 2018 Global Human Capital Trends report, organisations overwhelmingly agree that AI is very important. However, of the 72% of respondents acknowledging this, only 31% are ready to embrace it. Those HR teams that have already applied AI to their processes are reporting impressive results and setting the benchmark for an AI-led future.
Volume recruiters are already using AI to assess responses to screening questions. Using the internet, virtual interviewers can concurrently interview hundreds of applicants, narrowing the applicant pool to around 10%, helping assess prospective employees before opting for an in-person interview.
“The reduction in time to hire can provide a significant competitive edge, especially when competing for limited resources,” says Hnatiw.
“HR needs to acknowledge reality and find ways to embrace AI as part of the HR toolset” Darren Hnatiw, Chief Technology Officer, Frontier Software
AI is enabling passive candidate sourcing by identifying likely candidates from verified data sources. Using recruiter-defined parameters (eg location, skill set or industry), AI can return lists of potential candidates, predict their readiness for a move, develop and send communications, and map responses. Success rates for marketing campaigns based on these analyses have, anecdotally, been as high as 16%.
There are other workplace applications, too. Sentiment Analysis uses AI to examine the spoken or written word with the intent to identify the prevailing feeling or emotion. AI can compare individual employee computer activity – emails, browsing, etc. – to identify changes to defined baseline behaviours that might suggest an employee is considering leaving.
AI’s capacity for rapid analysis also lends itself to person-centric solutions. It can deploy lists of job-specifi c contacts for new starters, individual learning pathways and provide graphical representations of core competencies, comparing one employee to their peer group.
New employee self service systems are using chatbots to engage employees. Users are asked what they want to do, and the bot shows them how. For systems used infrequently by employees, chatbots enhance the user experience, providing an interactive assistant as the interface.
“One area that has attracted a great deal of positive attention is AI’s capacity to remove bias, particularly in recruitment processes,” says Hnatiw. “Using only data, AI isn’t influenced by writing style, candidate photographs, alumni or other affiliations. This leads to better and more diverse shortlisting. It has also been reported as selecting candidates suitable for horizontal career moves, via assessment of other, very different roles.”
For many HR practitioners, however, AI is a step too far. They view AI as something that removes the ‘human’ from the HR equation, and that may even hijack processes completely. A recent blog post by Ibrahim Diallo received worldwide attention when he described his own experience of an HR system automatically – and erroneously – terminating an employee. Due to the construction of the AI system, human intervention was unable to stop it from happening. The system sent termination advice emails throughout the organisation and the employee was forced to take three weeks’ unpaid leave until the AI was overruled. Aside from the inconvenience, potential legal ramifications and impact on earnings, the employee reported a change in his perceived trust relationship with co-workers when he was ‘reinstated’.
The relationship between AI and decision-making appears to be a focus of concern for those who don’t support it. Indeed, a 2015 Accenture survey found that only 14% of first-line managers and 46% of executives would trust advice derived from AI. Supporters of AI argue that, as human beings, we are not capable of comprehending the complexity behind the algorithm that draws a conclusion. Detractors counter with arguments around AI algorithms not being as nuanced as human thought processes. On a similar note, Hollywood and wider science fiction has not always painted a positive picture of AI either; most readers would be able to recall at least one AI-inspired movie that depicted a dystopian future ruled by robots.
Nonetheless, the reality is that AI is here to stay, not only within an organisational context but in every aspect of our lives.
“HR needs to acknowledge reality and find ways to embrace AI as part of the HR toolset,” says Hnatiw. “AI can supplement human interaction, but it’s not a complete replacement.”
Even with large-scale AI implementation, a human must still interview and select a successful candidate from a pool narrowed by AI. Data generated by AI can only be used to inform a decision made by a human HR practitioner. A human must confirm reported sentiment; humour, sarcasm and wit may present the exact opposite picture of the sender’s actual intention. Chatbots need a person to whom escalated transactions must go. Automated termination processes need input from HR to ensure the parameters are stringent and also reversible if necessary.
When viewing AI as more of a colleague than a cure-all, HR can assume a key role as AI evangelists, promoting and supporting company-wide applications. For managers, AI can provide succinct analyses of large data sets to enable data-driven decision-making. Managers can use AI to complete routine administrative tasks, such as planning and manual analysis, which is estimated to take up to 54% of their time. The hours recouped can then be invested in strategic and staff development activities that typically absorb about 17% of a manager’s time.
The impact on workforce composition and skill redevelopment that AI will bring is a future state that HR must begin to anticipate. With AI assuming responsibility for positions that perform repetitive tasks, the emergence of skill sets traditionally perceived as less important, such as design and creativity, will impact the future needs of organisations. Workforce planning models must also be updated to accommodate the possibility and impact of a future that includes AI.
For those who see the rise and importance of AI as somewhat exaggerated, remember this: the game-changing first iPhone was only released in 2007, sparking a smartphone revolution in its wake. Today, few people could do without them.
“Assuming that the pace of technological change is a constant is a mistake,” says Hnatiw. “It is exponential, not linear, so AI will leap forward significantly in a shorter time frame than that which brought us our smartphones.”
The human beings in winning organisations know this and are already preparing – make sure you are one of them.
This article was published in HRD Human Resources Director Magazine on 20 August 2018