Let's Chat Ethics

Is trust the make-or-break ingredient in the future of AI? This week we continue our exploration of the EU's Guidelines for Trustworthy AI and discuss the importance of trust in AI systems.

In the podcast we mention the work of the Z - Inspection led by Dr. Roberto V. Zicari from the Frankfurt Big Data Lab, Goethe University Frankfurt, Germany. Below we have linked the published AI Ethics Evaluation by James Brusseau, who is part of the research team for the AI medical research mentioned. 

• What a Philosopher Learned at an AI Ethics Evaluation

This week we discuss a paper where researchers detect moral values from people's writing. Is a Myers-Briggs test for morality the next step in employee vetting? 

Paper linked in article here 

Feliz Navidad everyone - our Christmas episode is here. Expect laughter, mince pies and mulled wine as we both reflect on the craziness of 2020 as well as the best and worst moments in ethical AI. 

BONUS EP! We have Ivana Bartoletti joining us for a compelling conversation about the use of our data and racist / sexist AI. 

Ivana Bartoletti is Technical Director, Privacy at Deloitte, and an internationally recognised thought leader in the field of responsible technology. Author of "An Artificial Revolution, on Power, Politics and AI", she starts a Visiting Policy Fellowship at Oxford University on international flows of data in January 2021, alongside her full time job. She is founder of the influential Women Leading in AI network. 

This week we chat about two hot topics in the news from the past two weeks.

• Can AI be the a matchmaking tool and enhance our love lives for the better? The news of Japan funding AI matchmaking to solve the declining birth rate. 

Following on from our transparency episode, where we spoke about the issues surrounding big tech funding ethical AI research - this week we discuss the controversial and wrongful firing of Timnit Gebru from Google.

Paper mentioned:

MIT Technology Review 

This week we discuss bias in tech that comes from the design of the tech itself rather than the bias that is automatically learned form data. 

Books mentioned:

Design Justice by Sasha Costanza-Chock

Invisible Women by Caroline Criado-Perez

This week we are back discussing the EU Guidelines for Trustworthy AI - looking at bias data sets. This is a two part episode where next week we will talk about bias by design. 

Should unethical research be published? As a peer reviewer should your own ethical values be considered? This week we discuss an ethics panel that took place at EMNLP 2020 - The 2020 Conference on Empirical Methods in Natural Language Processing.