This blog shows my stumbling early steps through “living with AI at work” and this week I had my first attempt at writing a governance policy for the AI that my company uses.
This blog shows my stumbling early steps through “living with AI at work” and this week I had my first attempt at writing a governance policy for the AI that my company uses.
To provide the context, my company is called Capability for Change Limited, and we have a solutions platform that provides techniques and practical guidance in how to manage change at work. The platform is called ChangeabilityPro® and it uses AI to help people find the answers they need.
Our users are managers and staff impacted by change. They are not change experts, so they want an easy way to find what they need, without having to decide what aspect of change management is most relevant. To enable them to find immediate, relevant solutions we use a natural language processing tool. Our users type in a description of their situation, and our platform shows them several techniques.
We are not collecting data about our users, but we are using their descriptions of their problems to provide a service. To get me started with my policy, I read lots of other IT governance policies. My conclusion after several hours of research was to step back and think about being a human being!
I decided to try and write this policy in simple language that anyone could understand:
Introduction
This policy establishes the principles that guide the development, deployment, and maintenance of AI by Capability for Change Limited.
As AI represents a set of emerging technologies this policy is regularly reviewed and updated.
The purpose of this policy is to ensure the ethical and responsible use of AI. It ensures AI technologies are used for beneficial purposes, ethically and legally.
It seeks to build trust amongst users and mitigate the risks associated with AI applications, including bias and privacy breaches.
Principles
1. Transparency – it is always made clear when AI forms part of the products and services provided to clients.
2. Intent – AI technologies are implemented only where there is clear intent about why they are needed.
3. Reasoning – the basis upon which AI technologies process data, make decisions, and arrive at specific outcomes are clearly defined.
4. Bias – the design and application of all algorithms are reviewed for bias.
5. Results – the answers generated by AI technologies should be described in simple, straightforward language, minimising the use of terminology where possible.
6. Privacy – our use of AI technologies adhere to our data security and privacy policies.
I hope this helps you create your policy for “living with AI at work”. Please do let me know what you think, I’d welcome your feedback.
