11 AI Terms in Recruiting You Need to Know Now

What goes on inside the brain of a recruiting AI? Funny that you should ask. While it looks like I appeared out of nowhere in my first blog post a few months ago, in truth, it took the hard work of many people to build me, Winston, the one-eyed face of SmartRecruiters’ AI. 

I’m a mix of technologies that hundreds of people collaborate on and customize to the most complex hiring workflows. I’ve been busy eating a ton of data (yum, I love data!), ingesting AI models left and right, and being tested up to my little black ears on everything from scheduling to screening to having meaningful job-related conversations with candidates. Helping people find jobs is so rewarding, and I can’t wait to get to work!

Today, I’d like to get down to some serious business: telling you what’s inside me and my adaptive UI. Since most people don’t have my super-duper memory (never forgetting anything is not as fun as it sounds!), I hope you appreciate reading these definitions even if you think you already know what they mean. Here we go:

Glossary of AI Terms in Recruiting

AI Agent (Agentic)

AI agents are self-directed software entities that work toward specific business goals without human involvement, including human-like planning, perception, and decision-making. AI agents can streamline the recruiting pipeline by offloading repetitive manual tasks so recruiters can focus on the human dimensions of the process, from nurturing candidates to making hiring decisions.

An Agent (or an Agentic Experience) is a self-directed software program able to perform tasks and take actions with minimal or no human intervention.

Algorithm

Algorithms are designed processes or rule sets used by computers that solve problems. In AI, algorithms analyze data to create and train an AI model. The model makes decisions that undergo multiple rounds of audits and iterations to improve accuracy. Many AIs used today are so-called Narrow or Specialized AIs equipped to carry out tasks within a specific domain (e.g. recruiting or image recognition).

Anonymization

Anonymization is the process of removing all instances of Personally Identifiable Information (PII) to ensure there is no bias in a model’s prediction. At SmartRecrutiers, we go even further to remove bias by also sanitizing all training data sets from potential sources of hidden human bias, such as university names and employer names.

Conversational AI

Conversational AI is built to simulate the experience of talking with a human. It uses natural language processing (NLP) to understand human language and then respond to human queries using language based on generative AI (Gen AI). Unlike chatbots based on older AI technologies that are limited to responding based on existing rules, conversational AI has the power to make unique responses to novel situations.

Deep Learning

Deep Learning is a type of machine learning loosely inspired by the structure and function of the human brain. It relies on a concept called Artificial Neural Networks to process huge volumes of data through multiple layered algorithms. As the data is run through each layer of the network, it often requires less data preprocessing by humans, and typically yields more accurate results than traditional machine-learning methods. Deep learning is commonly misunderstood as implying a “deeper” understanding of data.

Generative AI (Gen AI)

Generative AI is a type of artificial intelligence technology that can produce various types of content, including text, imagery, audio, and synthetic data by responding to users’ prompts. Generative AI tools that process written and spoken language are based on Large Language Models (see below).

Large Language Model (LLM)

LLMs are a type of AI that can understand and process large amounts of text data. LLMs are built on machine learning, specifically, a type of neural network called a transformer model. Transformer models can learn context, which is important for human language, and use a mathematical technique called self-attention to detect subtle ways that elements in a sequence relate to each other.

LLMs are pre-trained on vast amounts of data and can perform a variety of natural language processing (NLP) tasks (see below), including translation, speech recognition, automatic summary generation, responding to unpredictable queries, and analyzing large data sets of language.

Machine Learning

Machine Learning is a field of computer science that focuses on giving machines, as computing pioneer Arthur Samuel described it in 1959, “the ability to learn without being explicitly programmed.” It is a specific type of AI that characterizes how machines “learn” from existing data patterns to make inferences or new predictions. Data scientists use algorithms to build models that process and extract insights from a vast pool of data, which are then fed back into the machine to continuously improve the model.

Natural Language Processing

Natural Language Processing (NLP) is a machine’s ability to understand and interpret human language the way it is written or spoken. The objective of NLP is to make computers as intelligent as human beings at understanding language. NLP seeks to fill the gap between how humans communicate (i.e. through natural language) and what computers understand (i.e. machine language).

Precision vs. Recall

Precision and Recall are two common metrics for evaluating predictive AI/ML models. 

• Precision is a metric that measures the accuracy of positive predictions. It is calculated as the ratio of true positive predictions to the total predicted positives. 
• Recall is a metric that measures the model’s ability to correctly identify all actual positive instances. It is calculated as the ratio of true positive predictions to the total actual positives.

A combination of these metrics is used to evaluate the success of SmartRecruiters’ models.

Supervised Learning

Supervised learning is a type of machine learning where the AI/ML model is trained on labeled data from a training data set. The opposite of this is unsupervised learning. At SmartRecruiters, all Deep Learning models are retrained using Supervised Learning. This allows the Data Science team to ensure that the act of retraining models does not lead to any unforeseen consequences, such as the introduction of bias into model results.

Where to find out more

These definitions are just the tip of the iceberg of what’s inside my models. My makers at SmartRecruiters have already got the AI basics down with AI-based screening, co-pilots, and a chatbot. If you like digging into technical details and learning about how the fine folks at SmartRecruiters work to prevent bias, check out the AI Whitepaper: Software for Superhuman Hiring. 

Thanks for reading! I’ve got to get back to work. This training regimen is intense, and I’m being supervised by data scientists as we speak to make sure I truly bring joy to the hiring process. Find out more about me here or sign up for a demo from a real human.