How AI is Disrupting Job Boards, Recruitment & HR
<div class="grey-callout"><h2>In This Article You'll Learn</h2><p><ul><li>AI's Growing Role: Understand how advancements in AI technology are revolutionising the job board, recruitment, and HR sectors.</li><li>Impact on Employers: Explore how AI is aiding employers in tasks such as job description creation, interview scheduling, and candidate screening.</li><li>Implications for Jobseekers: Learn about the tools AI offers to jobseekers, from automated CV writing to real-time interview coaching, and the ethical concerns these raise.</li><li>Challenges and Ethical Considerations: Delve into the practical and ethical challenges that AI introduces into the recruitment process.</li></p></div>
Generative AI has been around for decades. It's simply a prompt > a model trained on data > and an output. But recently, there have been some incredible breakthroughs around software, the scale of training, and human reinforcement/feedback. It has caught people's imagination and will disrupt/transform many markets.
In this article, I'll explain how it will affect the job board, recruitment and HR market.
How AI is Currently Impacting Employers
Here are some novel ways AI is currently 'helping' hiring managers and HR:
Creating job descriptions
Many new AI solutions are designed to help employers create job descriptions. This can save time, potentially make them more comprehensive, reduce bias and stop common mistakes such as spelling and grammatical errors.
However, we should be aware:
- Job descriptions aren't that useful in recruitment.
- AI models have been unwittingly trained on lots of bad job descriptions, and as AI is then ironically trained on AI material, they will become the average of themselves.
Creating video job adverts and training videos with AI
Some people are getting (over) excited about creating video job adverts using AI. Traditionally, video job adverts have failed because the video production costs have been so high. Whilst I appreciate AI dramatically reduces these costs, there is little point if they aren't going to be an accurate representation of the company. Importantly, jobseekers don't seem that engaged: they like job traditional written adverts because it easy and quick to scan for relevant information, but this is more challenging with videos.
Where AI video creation has a lot of benefits will be training videos. You type a script, and an AI actor will 'present' it for you. It's straightforward to edit and make engaging content. Moreover, as training evolves, it is easy to update.
AI can be a recruiting assistant
This is where AI will deliver the most value as it can save a hiring team time by answering basic questions, scheduling interviews and asking basic screening questions.
However, it cannot match/rank applicants. More on this below.
Automated interview notes using AI
How often have you finished an interview, looked at your notes and seen you've only written a candidate's name? This can be solved using AI.
If you get the candidate's permission to record the meeting, AI can be useful for taking notes. But far beyond creating a transcript, it can summarise and categorise information. For example, explaining why a candidate is motivated for certain aspects of the job.
How AI is Impacting Jobseekers and Will Disrupt The Way Employers Recruit
A lot of the ways jobseekers are using AI can be considered cheating. Some may say it's just "using your initiative," but I don't accept this because a candidate is not authentic.
It will have an impact on recruitment...
AI can create professional 'photos'
The news is reporting about very clever AI software that can take approximately 20 photos of a person, then create about 100 'photos' of what that person would look like in a corporate environment. The individual often looks thinner and more youthful.
Whilst we shouldn't be biased, unconsciously, most of us are. So, it may have a minimal impact on initial screening. But it will be limited as many applicants don't include photos on applications (I don't recommend them), and ultimately the person is going to be interviewed so you'll see their true appearance - which I must remind you should not influence your recruitment decisions!
Ultimately I think it's a but of fun, like filters are on SnapChat videos.
Jobseekers are using AI to write CVs and cover letters
In the past, jobseekers have used CV writing services to create modern CVs that weren't an accurate representation of themselves. This wasn't too much of an issue, because the expensive CV writing services were easy to spot and few jobseekers used them. But AI has changed that; lots of applicants are now getting AI to write their CVs and cover letters.
This makes screening applications much harder because whilst we always accepted CVs weren't an accurate representation of a candidate, not they are more a complete work of fiction.
Just today I had three applications for the same job that were precisely the same! I rejected them all, and this isn't necessarily because I have a bias against AI, but because I knew they weren't genuine.
Whilst there are AI detectors, in our tests, they've been easy to spoof, and just recently the America Constitution was flagged as being written by AI!
Jobseekers are using AI to apply to jobs
Imagine this: a jobseeker uploads their CV to AI; the AI understands the jobseeker's skills and experience and then looks for relevant jobs on job sites by itself; the AI then applies on behalf of the jobseeker - writing a tailored cover letter and CV to that specific job, even answering the screening questions. Now, the applicant only has to wait for an interview!
This is happening right now. We are getting 'applications' using this technology.
You may be thinking these applicants are "lazy." But please appreciate applying for jobs is laborious, and most jobseekers never even get a reply for an employer (who they coincidentally may call "lazy".) Consequently, it is only a win-win for jobseekers to use this technology.
AI applications could be a nightmare for employers. They can't trust a compelling CV and cover letter or tell if the jobseeker was actually motivated to apply. Interestingly it could be a real challenge for employers using performance-based advertising because they are paying for 'jobseeker' clicks/views/applications, only it's not from a proper jobseeker.
It's also a nightmare for job sites. They want employers to get quality applications, and now they have issues detecting authenticity. There could be some difficult conversations between employers complaining about these AI applications, and the job sites may be on the receiving end.
Jobseekers can use AI to tell them what to say in video interviews
Video interviews have many benefits, but they may soon need an invigilator to ensure the applicant is not told what to say.
AI solutions currently exist that listen to a live video interview and coach the applicant on exactly what to say!
Jobseekers are using AI to prepare them for interview
Jobseekers have been preparing for interviews for a long time. They're used to reading books of "1001 Interview Questions & Answers", or watching YouTube videos from career counsellors etc.
The difference with AI is how much more detailed the advice is. Current AI solutions ask candidates to enter the job they're applying for and the company name. AI will then research the company, create some test questions and start interviewing the candidate! So when you ask a question in an actual interview, they've likely already rehearsed it and what your watching is more a performance.
At least when you ask "Why do you want to work for our company?" they'll have an answer!
AI is not going to help matching/ranking applicants (yet)
Some technologists believe AI will help accurately match/rank applicants, and they're working towards a future where AI solely makes the hiring decision. I accept the recruitment advertising model is not perfect, and whilst I commend their efforts, it is ridiculous.
With the greatest respect to a lot of the hiring managers I work with, many don't know what they want. Some reuse an outdated job description and may fudge it a little. They don't take the time necessary to create an accurate taxonomy of skills and experience that AI needs to match. It's the classic case of garbage in = garbage out.
AI created job descriptions are probably going to make this problem worse because so little consideration is given to a prompt; often its just "Act as an HR professional and write me a job description for a salesperson." In this example very little thought is given to how much marketing support the salesperson will be given, how long the sales cycle is, the value of each sale, is it telesales or face-to-face, how much account management vs new business is there, etc - all of which dramatically impact the job description.
The behaviour of hiring managers is unlikely to change because recruitment is often a distressed purchase. Normally a company is growing fast and they need someone fast. Or an employee has resigned, giving one month's notice, and they need to find someone in just one month! In most circumstances, there is little time to think: "Do I still need to hire for this role?" and if so "who do I really need?"
But the biggest issue is going to be the job application itself. Whilst we have some clever technology such as HR-XML and amazing taxonomies of skills and experience, how can we be confident it is accurate if it is all created by AI?
A technologist would explain how other marketplaces such as buying a property can be improved with AI matching. But this is much easier. Someone searching for a house (equivalent to the employer) can stipulate the property type; location; price range; number of bedrooms etc. Likewise, someone selling a property (equivalent to the jobseeker) can give accurate details. Both the person searching and the person selling are using known facts. But as already explained, many hiring managers need to explain what they want, and some jobseekers need to make authentic applications.
Another big limitation for AI is the data it is trained on. They suffer a very similar problem to psychometric tests (hence why I don't normally recommend them): no one is providing feedback to inform the models on whether the employee was any good! But the models need to know the equivalent of: this employee stayed with the company and was a Great Performer; or this person was a Poor Performer and left after one month. Otherwise, how can the models accurately predict the future?
There is also an ethical situation of AI making these important recruitment decisions. I recently saw an AI recruitment matching solution that video interviews a candidate and, during that time, calculates the candidate's BMI to determine if they would be too overweight for a role! This is unacceptable; "human resources" starts with "human" because it needs a human touch. Recruitment decisions are rarely binary; they are nuanced, requiring empathy and careful decision-making.
How do hiring managers solve these AI challenges?
AI is here and it isn't going away. As the saying goes, "the toothpaste is out of the tube, and it isn't going back in." So hiring managers and HR have to accept that whilst AI can be of some help, it will create challenges elsewhere.
We now know we can't give much credibility to covering letters, CVs, screening questions, video interviews and face-to-face interviews.
Both of these tools help us more accurately assess an applicant's background, skillset and cultural fit.
<div class="grey-callout"><h2><span class="text-color-purple">Key Takeaways</span></h2>
<p><ul><li>AI in Job Descriptions: AI can help create comprehensive and unbiased job descriptions, although the quality may vary.</li><li>Video Creation: AI can significantly reduce the cost of creating video job adverts and training videos.</li><li>AI as a Recruiting Assistant: AI can handle basic tasks like scheduling interviews and answering FAQs, saving time for the hiring team.</li><li>Automated Interview Notes: AI can transcribe and summarise interviews, aiding in the evaluation process.</li></ul></p>
<p><ul><li>AI-Generated Photos: AI can create professional-looking photos, potentially influencing initial screening.</li><li>AI-Written CVs: Many jobseekers are using AI to write CVs and cover letters, making it difficult for employers to gauge authenticity.</li><li>Automated Job Applications: AI can search for jobs and apply on behalf of the jobseeker, raising questions about the validity of the application process.</li><li>AI in Interview Prep: AI can provide highly tailored interview preparation, including real-time coaching during video interviews.</li></ul></p>
<p>Ethical and Practical Challenges:</p>
<p><ul><li>Authenticity Concerns: The use of AI in the recruitment process raises questions about the authenticity of applications.</li><li>Data Quality: Poorly defined job descriptions and lack of performance feedback can lead to ineffective AI matching.</li><li>Ethical Dilemmas: There are ethical concerns about AI making nuanced recruitment decisions, such as calculating a candidate's BMI.</li><li>Reference Calls and Job Simulations: These remain reliable tools for assessing a candidate's background, skillset, and cultural fit.</li></ul></p>
<span class="grey-callout">I want to thank James Neave who provided a helpful framework for understanding AI's use in recruitment, and Malcolm Myers for his input on recruitment marketplaces and the potential future of AI.</span>