The invisible AI costs

Artificial intelligence is changing how nonprofits work. For the social sector, it represents a huge opportunity. With limited resources and small teams, this tool is key to maximizing every donation, strengthening capabilities, and growing. 

But adopting AI isn't the same as being ready to use it. The 76% of social organizations already using it lack a strategy or usage policy. To leverage it, we need to be aware of its risks and use it with intention, responsibility, and strategy. 

The risks of using AI in the social sector

1. Inequality and bias: AI is not neutral. It learns from historical data and replicates the patterns it finds, including inequalities. If data is biased, AI amplifies it. Question results, monitor information quality, and involve communities from the design stage. 

2. Lack of transparency: 43% of nonprofits rely on a single person to make AI decisions. When there's no clear oversight, errors go unnoticed and risks increase. Make data usage, system functionality, and decision oversight visible.

3. Unprotected data: Working with AI involves working with data. And when that data belongs to vulnerable populations, the risk is greater. Protecting information means protecting people. Identify sensitive data, classify it, and anonymize it.

How to use AI responsibly?

1. Create a manifesto: It's the cultural framework that helps leaders and teams answer, "Why do we use AI?" Define when it's used, how it's used (and how it's not), and what principles guide its adoption. We share the Propel AI Manifesto as inspiration to build your own.

2. Define policies: Policies help the team use AI consistently and responsibly in their day-to-day work. Focus on three key moments:

• Before: Establish guidelines to check for errors and identify biases.

• During: Define specific roles for decision-making and reviews. You can also design alerts to detect errors. 

• Always: Document errors, share them with the team, and adjust processes when necessary. Adoption is a continuous improvement process.

3. Make it a habit: Using AI responsibly should be part of the team's culture. Before adopting any tool, ask yourself these five questions:

• Purpose: Does it support our mission?

• Data: Is it reliable, up-to-date, and free from bias?

• Transparency: Do we understand how it arrives at its results?

• Security: Have we tested it and can we detect errors?

• Oversight: Who reviews it and takes responsibility?

AI needs strategy and direction

The use of AI is not just a technological challenge. It's a leadership challenge.

The tools change; the purpose does not. The opportunity lies in creating the conditions for teams to use AI to multiply their impact. For that, we need clear rules, simple processes, and a shared vision.

Want to get started? We created this course to help you take the first step. Sign up here.

This blog was edited with support from AI. The creation is ours.