When you think of artificial intelligence (AI), what comes to mind? Do you imagine some sort of master computer that knows all? Do you think about films like I, Robot or Ex Machina?
Science fiction has warped our expectations of AI and robotics. We’re a far way off from a robot apocalypse, but AI has slowly entered into our everyday lives in a number of small ways.
Over the past decade in the recruitment industry, AI has played an instrumental role in sourcing and hiring candidates. AI systems have been reading resumes for years, getting increasingly accurate in doing so. These systems have allowed recruiters to skip lengthy database searches and go straight to the cream of the crop.
When reading resumes, AI applications make extensive use of natural language processing (NLP) and machine learning (ML), which may sound like a completely foreign language to some of you. Let’s explore both to understand just how the machines are reading resumes, what shortcomings they may have, and just how far this technology has come.
What Is Natural Language Process?
NLP a form of AI that allows a computer to understand, interpret, and sometimes manipulate human language.
Human language is complex. For example, “developer” could refer to a “software engineer,” “programmer,” “computer programmer,” or “coder.” These professions become even more specific when qualifiers like Java, Haskel, Solidity, and PHP are added.
To a person, especially one in the know, this all makes sense. Teaching a computer the subtle differences and intricacies, on the other hand, is a far more complex process. Furthermore, context has to come into play. If we were looking only at pure keyword matching, the following phrases would both be a 95 percent match:
“Does do Java development.”
“Doesn’t do Java development.”
However, in context, those sentences mean the complete opposite. This is something NLP can help AI understand, thereby increasing the accuracy of the matches it delivers when sourcing candidates.
Now, let’s place both of those statements in past tense. How does a system know to weigh past statements in relation to your skills today? An algorithm can sit in the back of the AI, identifying the most recent skills, weighing dates and longevity of those skills in an attempt to work out how senior a job seeker’s resume is. One of the key elements of any effective NLP is a good taxonomy, which can grow the knowledge base of the machine with the aid of machine learning.
What Is Machine Learning?
ML comes into play during many stages of recruitment. It is especially useful in matching job seekers to jobs, and it helps AI learn to make better matches over time.
For example, say a resume is uploaded into an AI application. Every day, the AI matches the resume with 20 potential jobs, and the job seeker selects the three most relevant matches of each day. Using machine learning, the AI will analyze these relevant matches, looking for any common structures, content, or language that indicates why these jobs are more relevant to the job seeker. Going forward, the AI uses the information in gained from these analyses to recommend more and more relevant matches.
What Does AI Think of Your Resume?
When reading a resume, AI looks for technical skills, soft skills, experience, dates, locations, and personal information. Common sense says the more often you list a certain skill on your resume, the higher the weight the AI will assign to that particular skill. The clearer you are with dates, the more accurately an AI system can measure your amount of experience with particular skills or roles.
These days, AI is so sophisticated that it can understand billions of words, sentences, phrases, institutions, qualifications, companies, schools, etc. The only real way for AI to misinterpret your resume is if the content of your resume is weak or inaccurate.
For anyone anxious about the effect an AI-enabled resume-scanning system will have on their job hunt, I have some simple advice: Focus your resume on the latest experience you have, and make sure to list within the document the sorts of keywords, sectors, and other phrases related to the skills you have today. Do not dwell on work experiences from more than five years ago. The majority of your resume should reflect your most relevant and current set of skills. If most of your resume is focused on what you did in college, you will be matched with roles that do not match your current level of experience.
AI has been part of recruitment for a while now, but tools like predictive matching, sentiment analysis, graph database technology, and more refined resume analysis are just on the horizon. Remember that whether its an AI or a human being, a decision-maker is only as good as the inputs it has. Bad information for the machine, bad result.
To make the most of AI recruitment technology, take the time to create the best resume possible. If you do, you will put yourself in line for far greater and more relevant job opportunities.
Arran James Stewart is the co-owner of blockchain recruitment platform Job.com.