Understand your impact

AI can save time, improve decisions and offer new services. But even simple tools can affect people in unexpected ways.

Before you use AI, check who it could impact and how. This helps you make better decisions, avoid harm and support your business goals.

Issues with AI can affect people in very different ways. For example, a chatbot giving the wrong information about clothing stock may be inconvenient. But the same error could harm someone seeking prescription medication.

AI tools can also create opportunities. For example, language translation tools can improve accessibility and help you reach more people.

As your organisation evolves, the impact of your AI may change too. Regularly check who may be affected so you can make informed and responsible decisions.  

Check your AI’s impact when:

• you start planning a new AI tool
• you expand how you are using AI  
• it might affect customers, your team or decisions
• it might influence hiring, pricing, access or eligibility
• you need to make a decision whether to use AI.

This doesn’t need to be a separate process. Build it into how you already assess risk and make decisions.  

Use this simple checklist:

1. Define the AI’s role

Explain what tasks it supports, what data it uses and what output it produces.

2. Identify and engage people who may be affected

Consider who it could affect, such as your teams, customers, suppliers, volunteers and the wider community. Pay attention to vulnerable or marginalised people in the community who can experience greater negative impacts of AI. 

3. Consider possible outcomes

Think about what could go wrong, who benefits and who may be disadvantaged.

4. Assess how serious it could be

Judge whether errors could be minor or serious, and whether they could be easy to fix or hard to undo.

5. Assess the risks

Consider the likelihood of each identified impact. 

6. Put safeguards in place

Reduce risks by assigning responsibility, checking AI outputs and monitoring performance over time by implementing the 6 essential practices.

Where potential impacts could be negative or significant, it’s important to pause and consider whether AI is the right fit.

Poor AI experiences can be frustrating. To maintain trust:

• provide a clear way for people to give feedback or challenge outcomes
• support customers and clients to contact a real person
• explain how AI informed or influenced decisions in plain language
• take action to address any harm caused if an AI system negatively impacts someone.

AI tools can have real-world impacts. They could shape customer or client experience, influence team decisions, and affect access to services. AI could also change how personal data is used and contribute to unwanted bias.

AI impact isn’t only about risk. It shows what’s working well, who benefits, and what trade‑offs you are making.  

Understanding impact helps you make clearer, more informed decisions about how you use AI.