We live in truly exciting times and when a (human) colleague of mine posed the question about unconscious bias of AI minute-takers, we agreed that even though in the realm of corporate governance, the accuracy and impartiality of board meeting minutes are paramount, the subtle influence of unconscious bias, whether human or AI-driven, can impact the crafting of these essential records.
As Company Secretaries and Corporate Governance Professionals we need to guard against our own unconscious bias. It is crucial that we strive to be factual, accurate and impartial when drafting these important (neutral and comprehensive) records which promotes good governance and accountability. As AI plays an increasingly pivotal role in transcribing and summarising meetings, it’s crucial to explore how unconscious bias can shape the narrative and unravel the nuances of these minutes, whilst also taking a moment to reflect on how to unveil and manage our own individual unconscious bias.
What can we do?
Understanding the differences between AI and human bias can help organisations make informed decisions about how to approach the recording process.
What is it?
Unconscious bias, whether in humans or in AI, refers to subtle, ingrained prejudices or preferences that affect decision-making or perceptions without conscious awareness. Unconscious bias in humans is a result of cognitive processes that occur without deliberate thought. These biases are often rooted in social, cultural, and personal experiences, and they can influence judgments, behaviours, and interactions. Unconscious bias in AI refers to biases that are inadvertently introduced into algorithms and machine learning models due to biased training data or design choices.
How to address it?
AI bias can be detected, analysed, and addressed through ongoing algorithmic improvements and ethical considerations. Human bias can be influenced by a range of subjective factors, and it may vary from one individual to another. Addressing human bias involves training minute-takers, raising awareness of potential biases, and implementing review processes.
Mitigation?
•Regularly assess AI-generated transcripts or summaries for accuracy, context, and potential biases.
•Train minute-takers to be aware of bias and provide clear guidelines for recording neutral and comprehensive minutes.
•Consider having multiple individuals review and validate minutes to reduce individual bias.
A combo of AI tech and human oversight can help mitigate biases and ensure that minutes accurately reflect discussions while minimising the risk of distortion or discrimination. Both human and AI bias underscore the importance of continuous vigilance, awareness, and proactive efforts to ensure fair and equitable decision-making and interactions.
Let’s be mindful and vigilant to ensure that the truth prevails!
At Aspida, the synergies of our people and technology are at the heart of what we do. Our team of experienced Company Secretaries and Corporate Governance Professionals continue to explore new ways of ensuring that we safely, securely and responsibly integrate and make use of cutting-edge technology such as Artificial Intelligence to provide the best outsourcing support services to our clients.
To find out more about our services and solutions, be sure to get in touch with us at info@aspidagroup.com
