Episode 7: Fundamentals of AI Ethics with Margaret Mitchell
I'm joined by one of the world's foremost AI/ML experts and one of the pioneers of the field in Margaret who founded and led Google's own AI Ethics division.
From medicine, social media, advertising, manufacturing: the pace at which Artificial Intelligence (AI) is eating our world means I couldn't help but get into the weeds into this topic and get someone in the pod who started thinking about this topic long before you, I and the general public caught wind of it.
Margaret has published 50+ blockbuster papers on vision-language and grounded language generation concentrating on the evolution of AI towards achieving positive goals in society. She is vastly famous for her work on the elimination of undesired biases concerning demographic groups from machine learning models. And her work is utilized by some of the biggest tech companies in the world. Margaret was also the founding member of Microsoft's ethical AI group.
In this episode, we talk about:
what is exactly AI?
what's machine learning vs deep learning?
the first principles of Ethical AI
various types of bias and the existence of a 'good' and 'bad' bias as key components to building an AI model
how governments and policymakers can evaluate ethical AI models
development of AI in the western world vs the Emerging markets
the possible utopian, dystopian and realistic predictions of a society fully adopting AI
You can listen to it by clicking on your favorite podcast platform of choice as below:
Here are the timestamps for the episode. On some podcast players, you should be able to click the timestamp for the episode.
00:00) - Intro
(06:03) - Margaret's background on Artificial Intelligence (AI)
(07:55) - Definition of Artificial Intelligence
(09:36) - Difference between Machine Learning & Deep Learning
(11:56) - Ethics in AI (Definition & Practices)
(13:43) - The role of the human layer in Ethical AI
(15:11) - Categories & Examples of biases that occur in AI
(22:29) - Normative vs Descriptive approach for selecting biases in machine learning models
(25:22) - Recent developments in the field of AI
(26:50) - Effective practices of ethical AI in big tech companies
(29:33) - Steps Governments & Policymakers can take to build and regulate AI models
(32:08) - How should tech startups in Emerging Markets develop models in the field of AI ethics?
(35:07) - Future of AI: Utopian vs Dystopian vision
(37:41) - Margaret's recent venture on open source AI
(40:36) - Outro
If you loved the episode, please recommend it to your family and friends!
You can also follow my twitter handle for more regular updates on everything innovation, civilization, history, and progress