Ethical AI Sourcing? What Does That Mean?
Hi everyone! Recently I got the chance to sit down and chat with the fabulous Karlee Porter about my new book, Digital Muse: Bringing AI Into Your Creative Process! One of the topics that came up was ethically sourcing AI creations. When we think about ethical sourcing, most of us picture fair-trade coffee or sustainably harvested wood. But as artificial intelligence increasingly integrates into every facet of our lives—from personalized playlists to healthcare diagnostics—it’s time to ask: What does ethical sourcing look like in the world of AI?
Behind the scenes of your favorite chatbot or image generator lies a complex supply chain of data, labor, and energy. Let’s explore what it means to ethically source the “materials” that power AI and why it matters more than ever.
Data as a Raw Material: Who Owns It?
Data is the foundational building block of AI—similar to how raw ore feeds the production of electronics. But unlike natural resources, data often originates from people: their photos, words, movements, and habits. Here are some questions you should be asking, and the best practices of organizations doing it right.
Key concerns:
- Consent: Was the data collected with full awareness and permission from its source?
- Representation: Are datasets diverse enough to avoid reinforcing bias and discrimination?
- Usage rights: Are creators of content (e.g. artists, writers) properly credited or compensated when their work trains AI?
Best practices:
- Use opt-in datasets where contributors are informed.
- Implement data minimization—collect only what’s necessary.
- Build transparent data audits into development cycles.
Invisible Labor: Who Trains the Machines?
AI doesn’t learn in a vacuum. When you hear news stories about AI and where it gets its data for training, you often hear about machine learning and “scraping” the internet. But AI has another under-reported learning mechanism. Much of AI’s “intelligence” comes from massive annotation and moderation efforts done by real people—often underpaid gig workers in the Global South—that enrich the data so AI can use it.
Key concerns:
- Fair wages: Are annotators and moderators being paid a living wage?
- Mental health: Are content moderators protected from exposure to traumatic material?
- Credit and attribution: Are these workers recognized as part of the AI creation process?
Ethical steps:
- Hire directly instead of through opaque third-party platforms.
- Offer transparent contracts, mental health support, and fair compensation.
- Advocate for industry-wide labor standards in AI development.
Environmental Costs: Where’s the Energy Coming From?
Training large language models or generative AI systems consumes enormous amounts of energy, sometimes equivalent to powering a small town. Many shy away from AI because of it. Yes, there are considerations, and we should hold these companies to higher standards.
Key concerns:
- Carbon footprint: How green is the data center where the model is trained?
- Water usage: Some AI infrastructure uses vast amounts of water for cooling servers.
- E-waste: What happens to outdated or discarded hardware?
Sustainable solutions:
- Partner with cloud providers committed to renewable energy.
- Measure and publish the carbon cost of training and inference.
- Invest in energy-efficient model architectures.
Ethics in Deployment: Who Benefits—and Who Doesn’t?
Even if the sourcing process is ethical, the application of AI must be considered. Who profits from these technologies? Are they accessible and equitable, or do they reinforce existing inequalities?
Considerations:
- Is the AI product accessible to marginalized communities?
- Are privacy and safety built into its design?
- Does it amplify harm (e.g., deepfakes, surveillance) or address real needs?
Accountability and Transparency: Who’s Responsible?
Ethical sourcing isn’t just about intentions—it’s about systems of accountability. Developers, companies, and users must ask hard questions and make space for oversight.
What helps:
- Open-sourcing models and datasets where possible.
- Clear documentation on data origin, labor conditions, and environmental impact.
- External audits and third-party ethics boards.
Conclusion
Ethical AI isn’t just about what we build—it’s about how we build it. Like any powerful tool, AI reflects the values of its creators. By thoughtfully sourcing its digital, human, and environmental inputs, we can push the industry toward a future that’s not just smarter, but fairer, cleaner, and more humane.