Fairness
AI must be designed to minimize bias and promote inclusive representation.
AI provides deeper insight into our personal lives when interacting with our sensitive data. As humans are inherently vulnerable to biases, and are responsible for building AI, there are chances for human bias to be embedded in the systems we create. It is the role of a responsible team to minimize algorithmic bias through ongoing research and data collection which is representative of a diverse population.
01
Real-time analysis of AI brings to light both intentional and
unintentional biases. When bias in data becomes apparent, the
team must investigate and understand where it originated and how
it can be mitigated.
02
Design and develop without intentional biases and schedule team
reviews to avoid unintentional biases. Unintentional biases can
include stereotyping, confirmation bias, and sunk cost bias.
03
Instill a feedback mechanism or open dialogue with users to raise awareness of user-identified biases or issues. e.g. Woebot asks “Let me know what you think,” after suggesting a link.
“By progressing new ethical frameworks for AI and thinking critically about the quality of our datasets and how humans perceive and work with AI, we can accelerate the [AI] field in a way that will benefit everyone. IBM believes that [AI] actually holds the keys to mitigating bias out of AI systems – and offers an unprecedented opportunity to shed light on the existing biases we hold as humans.“
To consider
- Diverse teams help to represent a wider variation of experiences to minimize bias. Embrace team members of different ages, ethnicities, genders, educational disciplines, and cultural perspectives.
- Your AI may be susceptible to different types of bias based on the type of data it ingests. Monitor training and results in order to quickly respond to issues. Test early and often.
Questions for your team
- How can we identify and audit unintentional biases that we run into during the design and development of our AI?
- The status quo changes over time. How do we instill methods to reflect that change in our ongoing data collection?
- How do we best collect feedback from users in order to correct unintentional bias in design or decision-making?
Fairness example
- After sitting down with members of the hotel’s global management, the team uncovers that diversity and inclusiveness are important elements to the hotel’s values. As a result, the team ensures that the data collected about a user’s race, gender, etc. in combination with their usage of the AI, will not be used to market to or exclude certain demographics.
- The team inherited a set of data about guests from the hotel. After analyzing this data and implementing it into a build of the agent, they realize that it has a degree of algorithmic bias from the data. The team proceeds to take the time to train the model further on a bigger, more diverse set of data.
