As AI adoption has ramped up, the uses of AI have moved from theory to practice. That means we’re now coming face to face with the reality of AI, and its ethical implications.
It’s increasingly important to figure out the consequences – intended or unintended – of the AI you implement. And it’s even more important to make sure your AI reflects the values that your company holds. So how can you do that?
Start With Your Values
To use ourselves as an example: our three core values are accuracy, empathy, and utility. In practice, that means:
- Accuracy: Everything we say has to be true, and support the trust we build with our clients.
- Empathy: We prioritize our clients’ success and are dedicated to understanding what they need to make their lives better.
- Utility: Everything we create for our clients has to be actually useful.
When it comes to choosing and building AI for our own use, those values are always at the forefront. And with those values in mind, we have some ethical considerations that we prioritize for our AI as we create it:
- Transparency: We need to know everything that’s going into our AI systems, how they operate, and where answers are coming from.
- Privacy & security: Our AI must protect our clients’ data, and operate in a safe way.
- Reliability: The AI tools we build have to work, and work across contexts.
There are some other important ethical considerations to be taken as well. There needs to be a chain of accountability for all AI tools, with humans at the root of it. And we want our tools to be fair, and to not perpetuate discrimination or inherent biases. All of this matters to ensure that our AI technology is contributing to the kind of world we want to live in.
Other Examples We’ve Encountered in the Wild
- One of our clients is a large music company that represents over a thousand artists. For them, human creativity is a foundational company value. As such, they don’t want to use AI to create images or replace artists’ work; they want to focus on data-driven applications.
- A CPG company we’ve worked with has invested heavily in saving resources by streamlining their packaging, which results in both cost savings and a positive environmental impact. When we helped them consider AI options, we made sure to focus on solutions that were both energy-efficient and cost-effective.
So How Can You Make Sure That Your AI Implementation Supports Your Values?
- Identify your organization’s values to begin with. (Microsoft put together a great rundown on some important principles, that may help you clarify where to put your attention).
- Come up with a list of non-negotiable ways that your AI needs to meet those values. What are the qualities your AI needs to have to live up to your company image?
- Hold regular check-ins to determine how things are going, and whether you need a course correction.
Taking the impact of AI seriously leads to thoughtful, interesting, and effective implementation. It’s a strategy that’s better for businesses, customers, the planet, and the culture we’re building around AI. We’re at the beginning of a big wave of change, and we have the power to determine whether AI will make the world better or worse by the way we implement this new technology. It’s something we should take seriously.