The AI Revolution in Recruitment: Balancing Efficiency and Human Touch

AI in recruitment - will we see an AI arms race in hiring?

Although more and more employers and job candidates are using AI in the recruitment process, both sides appear to have concerns about the other's use of the technology. For instance, 58% of job seekers around the world are using AI tools to help them in their job search, according to a survey by Capterra. But 83% of this number admit to having used the technology to exaggerate or lie about their skills. Rebecca Carr is Chief Executive of recruitment platform provider SmartRecruiters. She suspects that even more people could be using generative AI than the study indicates: I bet the number is higher - that's my hunch. But it depends on how you define 'using AI'. Certainly, it's very likely that a lot of people are using it to help them rewrite the description of themselves on their resume to make it more attractive. But I don't know if they've been included in the figures. Carr is also not convinced that fraudulent activity is as widespread as the figures would suggest: I don't think lots of people are 'lying' in the way employers fear. For example, a recruiter would say they have good sales skills as they're selling jobs, but that doesn't include the quota-carrying skills you'd be looking for in a salesperson. So, are they lying or just enhancing their resume? I'd say that people have been 'lying' in the job application process for a long time. While AI may make ways of doing so more accessible, that's not to say employers haven't been duped in the past, just possibly that enhancement is taking place in higher volumes now. Peter Wood, Chief Strategy Officer at recruitment agency Spectrum Search, agrees: I don't think candidates exaggerating or lying about their skills is as common as people think. A poorly written AI-based CV is quite easy to detect, although less so if candidates use their initial CV and apply prompts to it to raise the temperature on their key skills. But they often have those skills in place. They just don't necessarily know how to present them very well or aren't conscious of the attributes a hiring manager is looking for. So, you could argue they're using Generative AI to help them leverage their skills in a better way. However, says Carr, there are certain common indicators for employers that candidates may have been using generative AI: There are certain words that the average human doesn't use a lot, which show up in resumes due to the way the algorithm is trained by the people who first created them. Although it depends on the program, these kinds of words include to 'delve' into a problem, 'embark' on something, or say something is a 'testament' to their skills. Some employers look up the quantity of these words to see if a resume has been augmented, and bots can sometimes come up with a likelihood score. But it's often hard to say for definite. Alex Thornton, Talent Acquisition Lead at the Craft Distilling Group, estimates that between 30% and 40% of job seekers applying to his business currently use AI in some form during the recruitment process. But it does vary: Candidates might only use it to structure things better, but it does tend to be younger people who do so. I'm also seeing people using it more as the job market slows because they want to find a way to stand out. The problem with using AI is that CVs all look the same. So, while people have my empathy, it's not great as it has a homogenizing effect. More senior hires don't usually use it as the companies they've worked for and their length of service speaks for itself, and they don't need to sell themselves in the same way. Another challenge for employers posed by AI, according to the Capterra survey, is that over a quarter of candidates also use it to mass-apply for jobs, potentially clogging up employers' recruiting pipelines with undifferentiated applications. The problem is that this adds to their hiring costs because of the amount of time and effort required to wade through endless resumes. But this scenario inevitably raises the specter of a potential AI arms race, in which employers implement more and more AI applications, including AI detection tools, to counter the growing use of the technology among job seekers. As Carr points out: Where an arms race could become most exaggerated is if companies start receiving lots more job applications than before because AI applies widely for jobs on applicants' behalf. Screening and getting through loads of resumes could force employers to use more AI as having a person do it involves a much higher cost. But Wood isn't convinced: I don't expect to see any great change in the next few years. HR managers often aren't very tech-savvy and they need buy-in to purchase additional tooling to do their job better. So, they need sign off, there's the lead time, and the cost involved too. But I suspect the majority aren't seeing this as a huge problem just yet. I also don't expect to see an AI arms race in future as the recruitment process is about more than just a CV. Thornton agrees. Although he understands why some recruiters use AI to save time at the screening stage, he had a bad experience with the technology in the past when one job applicant was assigned a zero percent match. Luckily, their resume caught his eye and on digging a bit deeper, he found they were actually a great candidate. As a result, Thornton now prefers to assess CVs himself as he does not believe the technology is sophisticated enough to do the job effectively. He explains: CVs don't always tell you everything. Someone might have the right expertise but not include a super-important skill as they don't feel it matches the job description. My experience as a recruiter means I'll pick up on that, but AI won't. I possibly do a lot more screening calls than other people, which takes time, but I think it's worth it because I want to give everyone a chance. Since I joined the company six months ago, I've hired 27 people, from junior level to the C-suite, and my attrition rate is zero. The second part of this two part-series will look at how the technology is likely to change the hiring process, and the role of recruiters.

AI in recruitment - will the technology lead to a change in the hiring process?

The first part in our two-part series on 'AI in recruitment' assessed whether there would likely be an AI arms race in the market. In this second part we explore how the technology is likely to change the hiring process, and the role of recruiters. Part one looked at whether we are likely to see an AI arms race manifesting in the hiring process. Alex Thornton, Talent Acquisition Lead at the Craft Distilling Group, uses AI to undertake some admin work. But he is unsurprised to learn that three out of five job seekers would be more likely to apply for a job if they could interact with a human at any stage of the process. The same Capterra survey also reveals that 38% of job seekers would reject offers from companies using AI-heavy hiring processes in which human interaction was minimized. As Thornton says: People want to talk to people. They want the human connection. It gives them the opportunity to sell themselves and you can't do that with AI. It doesn't pick up on charm or charisma, just skills, and people don't like that. One of the biggest things is organizational culture and whether people will get on, and you can't use AI to assess that. Peter Wood, Chief Strategy Officer at recruitment agency Spectrum Search, agrees: I don't believe that any hiring process should be 100% automated - ever. There'll always be a need for human interaction. This is a very personal and human process, so humans need to be involved in it. People aren't ready for AI-based interviews, for example. So, while the technology may be there, I'm not sure we as humans are, and won't be for another 10 years minimum. How I use AI internally is to supercharge others, so it's never about replacing them, just helping them do their job better. And at the front end, it's always about human interaction. In the experience of Rebecca Carr, Chief Executive of recruitment platform provider SmartRecruiters, meanwhile, senior hires still expect to interact at every level with people rather than bots. But the same is not true of more junior or high turnover roles, where "people are OK with it". However, she adds: I'd say it's the people in the middle who are getting most frustrated with the automation of the recruitment process. So, it's likely to be people with degrees or specific skills in roles, such as project management, who are most unhappy and find their expectations aren't meeting reality. However, she believes that AI has a role to play in three key areas of the hiring process. These are: Here AI has to predict what you want personally as a candidate or hiring manager, so it's not just about what skills are available. It's wildly personal, so there'll always be a need for subjective assessment. Where it might dwindle is with gig work as more jobs move that way. Uber drivers, for example, aren't interviewed by Uber. It's binary, so are they good or not, and do they have a driver's licence or not? It's supply and demand. But if people have to go into an office and interact with others, there'll always be a need for conversations and assessments, although AI may make recommendations. As a result, Carr believes that 'efficiency' and 'insight' will be the first categories to be automated, while 'discovery' will be "harder to get right". This is because, on the one hand, there is much inconsistency and subjectivity in classifying how people work, which means no standard definitions currently exist. On the other, much HR data, including employee performance information, is not clean, which makes it difficult to train AI models. As Carr points out: It'll work in pockets and in some jobs. For example, you could read a programmer's code and get an objective assessment of their quality. But it's much harder to get an objective assessment of a CEO. You could use the business results, but how do you know if a good or bad performance was down to them or their chief of staff? So, it's challenging. If you could figure out how to do it, it'd be very powerful. But it'll be the hardest to do, and it's also the most highly regulated area, so there'll be roadblocks in terms of how you store and access data too. Nonetheless, says Carr, we are still very much in the early days of deploying AI in this space: Blackberrys, if you remember them, were a stepping stone to smart phones. And I'd say we're now in the Blackberry phase of AI in the hiring process. So, it was a little clunky and your fingers were often too big to hit the buttons, but it became part of how you operated over time - and I think AI will be the same. Does a smartphone entirely take away from interacting with someone? No, although we might see more conversations out of the usual workflow due to texting. But people still want to connect. A big piece of hiring in future may be tech-based, but we're always going to want to speak to someone sometimes to assess them and make more subjective decisions. Within five to seven years though, Carr expects the recruitment process to have totally changed. She explains: Most of the world will hire and be hired differently. All admin work will be fully automated, and as a candidate, jobs will be presented to you as an incredible match, or you'll at least be able to find them very quickly. You'll apply and be assessed by agents and apart from one to three conversations with humans, the rest will be automated. So, the process will be almost entirely led by agents, including onboarding, which will be fully automated. As part of this shift, Carr believes the role of a recruiter will become much more strategic and much less admin-based: A big part of that will be focusing on talent and organizational design. It'll be less about sourcing candidates and scheduling interviews and more about figuring out how the business gets in front of problems and hires the right skills to get there. There's a significant labour shortage on the horizon and jobs will have to be automated to deal with it. The birth rate is declining dramatically, which implies more jobs will need to be done by agents so that people can focus on learning, research and decisioning. This means recruiters will ultimately become strategists for talent. The key message here is that no matter how useful AI might be at boosting the efficiency of the recruitment process, people still want to deal with people. So, even if an AI arms race does come to pass, the secret to success on both sides will always be the human touch.