Measuring the impact of our software
At Tengai, we recognize that our software can have an impact on individuals and society as a whole. We take this responsibility seriously. Because without the right safeguards, AI could cause harm and exacerbate existing inequalities. One critical aspect of responsible AI development is, therefore, the focus on identifying and mitigating bias. As our society continues to evolve with rapid innovation in emerging technologies, in particular AI, we will continue to evolve and further develop our principles to ensure that we hold ourselves to the highest standards.
Tengai AI ethical principles
The following AI ethics principles work to guide us as we develop our artificial intelligence (AI) technology.
1. Be beneficial to society
We recognize the real-world effect our AI can have and we are committed to developing AI that will be beneficial to both individuals and society as a whole. We will develop our AI to help our customers identify people with the right competencies for the role at hand and increase the quality of hire in every process. Our goal is that our AI will be is used to actively promote diversity and aid in the achievement of equal opportunity for everyone regardless of gender, ethnicity, age, or disability status.
2. Avoid creating or reinforcing unfair bias
We implement several bias mitigation strategies to ensure that our AI will not reflect or reinforce existing inequalities. We will design our AI to not hold any preconceived or subjective opinions toward a candidate to avoid an unjust impact on people. We will strive to build diverse teams, as we believe representation is crucial to examining our AI in a more critical way.
3. Be designed to help people make better decisions
Our goal is to empower human beings, allowing them to make informed decisions and fostering their fundamental rights. We will develop our AI to objectively score competence and personality traits related to work performance. Our AI will assist recruiters with unbiased candidate data and help HR professionals make better decisions.
4. Incorporate privacy design principles
We will continue to incorporate privacy and data protection at each step of our technological development and deployment. We will design to encourage a foundation of privacy safeguards and provide appropriate transparency and control over the use of data consistent with best practices and legal standards. We will store and secure candidate data in data centers on Swedish soil, honoring GDPR and strict Swedish data storage laws.
5. Uphold high standards of scientific excellence
We will aspire to high standards of scientific excellence as we work to progress the use of AI in recruitment. We will work with psychometric experts to validate our AI and draw on multidisciplinary approaches. We will continuously do internal tests to ensure that our AI is objective and aim to have experts validate the technology on a yearly basis.
6. Respect the law and act with integrity
We will employ AI in a manner that respects human dignity, rights, and freedoms. Our use of AI will fully comply with applicable legal authorities and with policies and procedures that protect privacy, civil rights, and civil liberties.
"The validation study implies that Tengai accurately can assess personality traits correlated with work performance and be used to achieve a more unbiased recruitment process!"
Dr. Anders Sjöberg, Psychometrics Sweden AB
How Tengai works to prevent and mitigate bias in our AI
Our mission is not just to avoid bias in employment decisions, but that our AI is used to actively promote diversity and aid in the achievement of equal opportunity for everyone regardless of gender, ethnicity, age, or disability status. While our goal is to amplify, augment, empower, and enhance human performance, we recognize that there already are existing power imbalances in society that need to be considered.
Bias mitigation strategies
To ensure a fair AI without algorithmic bias, we have developed several bias mitigation strategies. One strategy we have had from the beginning is to never use face recognition software with our AI to take away any chance of possible discrimination based on appearance. In addition to this, we developed our AI without any access to candidates' personal, identifiable data and made the interview process exactly the same for every candidate in each specific process.
Our AI is interactive in its nature, meaning that we encourage a conversation between the candidate and the AI and use natural language understanding (NLU) which gives it the ability to understand human language as it is spoken. The bi-directional conversation is possible through our NLU that we constantly tweak to pick up nuances in language depending on demographic. Saying "Yes" and "Aaarg" (if you're a pirate) is equally valid, one is just more common than the other. The interactive aspect of our AI is what creates a fun candidate experience. But to create a bias-free assessment, candidates' answers are always processed through a scientifically proven framework that ensures that only the quality of the answers is taken into account. Our AI never considers the length of the answers or the language quality - meaning that just because the candidate has an accent in a foreign language does not mean they should be treated differently.
A final example of a bias mitigation strategy we implement at Tengai is having a diverse team with individuals from different backgrounds, ages, and gender identities. We believe that representation and having a diverse development team are crucial to mitigating subconscious stereotypes and not getting a collective blind spot. By allowing different perspectives, we can examine our AI in a more critical way and examine how bias could influence our AI.
Additional privacy policies and procedures
Data Security & Privacy Statement explains how we process, gather, use, store, share, secure, retain, and dispose of Confidential Information and Personal Data on behalf of our customers and their users.
Are you curious to learn more about Tengai?
Book your own walk-through demo.