Balancing breakthroughs and ethical boundaries with data and AI

How do we deploy AI solutions without doing more harm than good? This question should run through the mind of every AI practitioner whenever they deploy a solution. In this talk, I’ll look at some of the hidden costs of the AI revolution. The under-represented groups, ignored by biased training data. The use of exploited workers, labelling data, or taking the place of AI systems when the system isn't quite ready. The artists, musicians and writers who have their work scraped and fed as training data into these enormous foundational models.

Finally, the environmental cost. With the power usage of OpenAI predicted to grow to exceed the entire energy consumption of Argentina by 2027, AI could be a climate disaster. However, there is a route through this maze of ethics and responsibilities. One that allows us to take advantage of the incredible benefits of AI. One that has the potential to transform public services for the better, free up resources, improve reliability, and even save lives. In this talk, I’ll propose a considered approach to adopting AI.

By evaluating its impact on the broader ecosystem, rather than just the immediate and apparent effects, we can harness this incredible technology to deliver truly revolutionary transformation.