I am equal parts fascinated and cautious of the customer benefits that are being realised everyday through the constantly evolving application of AI related techniques like NLP, computer vision and deep learning.
When applied poorly all sorts of unintended outcomes are possible.
Fairness and ethics in AI and the challenges in designing and deploying a useful machine learning models all share one common factor for success - good data.
Good data is hard to come by, often locked across institutional, commercial and individual privacy bounds. Improving bad data is often hopelessly expensive and difficult.
My time over the last few years has been spent with tech vendors, researchers and open source communities finding better ways to build customer-oriented, secure, privacy-aware, made-for-machine-learning data pipelines and capabilities. I'm also part of the UK All-Party Parliamentary Group for Longevity, you can read our latest report here.
I've got to spend a lot of time deep diving into new privacy theories like Prof. Helen Nissenbaum's Contextual Integrity and emerging tech like federated learning, differential privacy, homomorphic encryption and secure multiparty compute. To see my experience check out my LinkedIn.
I've witnessed the challenges and creative de-risking solutions that organisations are taking when signing data exchange agreements.
Today, I am working with a great team to create new ways of establishing trust in data sharing ecosystems at Flusso.
I'd love to talk about how these new ways of doing things can help your business and your customers. Please feel free to contact me to chat about any of the above!