Pelumi Obasaju
At Oxford University’s first women’s college, Lady Margaret Hall, some of the most eminent women working in AI ethics gathered from all over the world. September 16th 2019 saw the materialisation of the first ‘100+ Brilliant Women in AI & Ethics’ conference; inspired by a list put together by Mia Dand in autumn 2018.
The ORBIT organised event allowed those in attendance to participate in discussions about AI governance, diversity and society. This event had collaboration and diversity at its very heart, with keynote speakers and workshops that looked to the future and explored how AI can work in an ethical way with and for society.
Uncertainty, Diversity and Responsibility
The first keynote by Jaquelyn Krones of Microsoft explored “The richness of uncertainty” when moving from AI ethics principles to practice. She focused on how fairness is not a checklist as it may be in conflict with other principles such as privacy and security or reliability and safety. However, this “cross-discipline productive friction” is not only critical to human creativity but is needed to engage stakeholders in solving thorny AI development problems.
She helped set the tone for the day by challenging how we talk about AI and reminding us that the technology is not the driver of society; it is us that design AI and we have the power to make the decisions and choices.
Delegates then broke out for various workshop sessions: Algorithms and society, AI & Global Governance, Data and Decision-making, and The World of Work. Each composed of a morning and afternoon session, followed by a plenary panel on the work done in each workshop. The outputs, where appropriate, will be published.
Dame Wendy Hall delivered the second keynote on “Socially Responsible AI”. Her talk highlighted how the growth of jobs is linked to diversity, and the importance of diversity as part of any ethical framework. She honestly summarised that “if it’s not diverse then it’s unethical”.
The diversity thread was woven through Beena Ammanath’s keynote on “Diversity- The AI Success Imperative”. This call to action echoed that the time for diversity is now, whilst beautifully illustrating the nuances of diversity.
Language and Our Future
In conversation with writer, Jeanette Winterson, and Baroness Beeban Kidron raised some interesting points. Jeanette Winterson emphasised the power of language, challenging the jargon in the conversation around AI and urging us to think on the words we use as “the language creates the thought”. She also led us to consider the importance of how we educate AI machines and develop algorithms, drawing insights from Mary Shelley’s Frankenstein to avoid the mistakes of the past leap-frogging themselves into the future.
Baroness Kidron’s keynote on “Overselling AI: The Problems of an Automated Childhood” addressed her work protecting children on the internet. She definitely left a lot of food for thought, querying if childhood should be reconceived for AI and emphasising that the rights of children should be coded into AI. She uttered a particularly harrowing line: “Silicon Valley said move fast and break things…but we did not consider that the thing that may be broken is our children “
The excitement, enthusiasm and energy of this conference cannot be fully captured in a summary. It was refreshing seeing so many women who are working to ensure that the benefits of technology are for all in society. These women are leading the way in industry, professional services, academia, policy and civil society.
It’s inspiring to see the impact “simply starting a list” has. As Mia Dand’s vision expands, encouraging and supporting more “unreasonable women” who change the world and recognise more women doing work in AI and ethics that were not on the original list. Hopefully, more women will begin to occupy space in every level of AI. Women need to be part of the conversation regarding the future because the future affects us all. Women are in the “all”.