MELBOURNE, Australia, Sept. 7, 2022 /PRNewswire/ — Sapia Labs, the research and development division of global AI company Sapia.ai, has released innovative research highlighting the effectiveness of Natural Language Processing (NLP) in accurately inferring personality traits through text-based structured interviews.

The study, titled Explainable Personality Prediction Using Answers to Open-Ended Interview Questions, positions Sapia at the forefront of advancements in NLP-based assessments. It reveals a significant average correlation of r=0.37 (p < 0.001) across the six HEXACO personality dimensions, indicating that Sapia’s NLP technology can effectively infer personality traits from language.

The accuracy of Sapia Labs’ findings is attributed to the integration of its proprietary dataset with Google’s BERT (Bidirectional Encoder Representations from Transformers), the leading algorithm in natural language pre-training. This refined version, named InterviewBERT, draws from a robust dataset of 3 million interview responses and 330 million words, collected from a diverse sample of 47 countries. The resulting algorithm excels at capturing the contextual meaning of words in interview replies to infer personality traits.

Dr. Buddhi Jayatilleke, Chief Data Scientist at Sapia, emphasized the significance of the research: “Our technology makes personality assessments more accessible, mitigating the subjective biases often present in human interviewer evaluations.” He noted that written interview responses encode signals that predict job fit, leading to a better candidate experience, particularly with the positive feedback received regarding the asynchronous chat format.

Barb Hyman, CEO and founder of Sapia, highlighted the profound impact of this technology for those looking to utilize AI tools grounded in ethical practices. “Our AI’s ability to interpret the contextual meaning of words in job interviews and accurately infer personality traits in an explainable manner establishes us as the leading chat-based interviewer on the market,” Hyman stated.

She further explained that Sapia has developed a fairer assessment method, free from the biases associated with human judgment, ensuring consistency across all candidates. “Even anonymized CV data often contains biased information related to education, experience, and gender,” Hyman added. “This is why many trusted consumer brands rely on our AI technology to identify the best talent without bias.”

In addition to its NLP advancements, the study addresses Explainable Artificial Intelligence (XAI), a methodology that allows users to understand and trust AI outputs. By utilizing natural language, the transparency of the interview responses is enhanced, making them easier to interpret compared to other assessment methods, such as video or game-based AI.

Sapia’s innovative technology is already in use by some of the world’s most trusted consumer brands, including the Qantas Group and Woolworths Group.