Eight simple rules for using AI at work

There’s been a lot written about AI, the efficiencies it brings and how it will transform the workplace. In some cases, new tech has already been implemented – three quarters of charities are using AI, according to the 2025 Charity Digital Skills report. 

Despite the heralded advances, it’s not a free-for-all. In this article, we share some simple rules and best practices to steer how you use AI and what to take precautions against.  

How to use AI responsibly at work 

Rule 1: Know what you’re using AI for 

If you’re going to use AI at work, understanding its use-cases, scope, and limitations is important. From an organisational perspective, drafting an AI governance policy sets out the standards which staff should abide by.  

Some concepts to include in an AI policy document are ethical guidelines, examples of when to and not to use AI, along with who is responsible for owning the technology within the charity.  

Rule 2: Mitigate against bias risks 

There are many considerations around how AI algorithms select appropriate reference datapoints.  Data could be gathered from around the internet or from a subset of select sources. Understanding where the information comes from may uncover unintended biases.  

For example, AI applicant tracking systems may determine that certain backgrounds have higher hiring success rates. It might conclude that it makes logical sense to promote candidates from those origins. Bias is inherent here – that’s not a fair policy for selection. To mitigate such risks, IBM recommends using AI fairness tools to evaluate.  

Rule 3: Share how and why you’re using AI 

There are efficiencies to be gained from AI in the workplace. In finance, the technology  could assist directors in automating mundane reconciliation tasks and discerning trends. Meanwhile, in the fundraising team, AI could help speed up the time to produce a grant application.  

AI could also be used to engage audiences via chatbots and other service delivery automatons. Audiences may want to, or have the right to know when they are dealing with Al. It’s up to charity leaders to decide how and in what format audiences are told.  

Rule 4: Test your AI and study its results 

Diving into the AI assistant, charities should be cautious. Due diligence should be performed on the assistant in terms of what it’s meant to do (i.e. answer questions or something more?); who it speaks to; the tone of language it uses; and how it is implemented.  

Charites should proceed only when the AI is tightly controlled and tested. Leaders will want to ensure that AI bots have a firmly circumscribed role so that there’s no added moral or ethical liability.  

Rule 5: Check and check again