World Economic Forum, 22 Jan: The day started with a great “Human + machine” conversation about AI over breakfast, upshot we need to redesign (or reimagine) business based on AI capabilities, reskill workers to leverage their existing knowledge to combine with AI to produce the new roles of the future and be responsible through accountability, transparency, honesty, fairness (inclusion) and human centric design.
I spent an insightful 90 minutes with many of the founders of the World Economic Forum's Cybersecurity Centre, which Zurich is a partner of. There we discussed the goals of centre to reduce vulnerability, contain attacks and restrain attackers including key issues around how to close the resources gap, need for public private collaboration and many aspects of data sharing. Then a short walk to the future of insurance where we debated what topics we should work on an as industry supported by the Forum, including a diverse range of thought leadership on the role and ethics of AI and how to reskill workforce, protection gaps, climate change and sustainability more broadly, ageing healthier, mental wellbeing...plenty to do in 2019!
I heard from Bloomberg, BoA and Goldman Sachs on diversity and equality, with the call for us to all do more...a challenge from Citi for us to not talk ourselves into recession...and from the founders of water.org on their efforts to provide clean water to the millions who don’t have access to it.
The afternoon was spent in conversations about the use of customer data and AI in financial services discussing the kinds of principles we should be using, such as; fairness, explainability of use or results, accountability for decisions, transparency. Ultimately the acid test is the need for acceptability of the outcomes, through minimising bias and social exclusion. A few other relevant areas we touched on were the need to be able to explain third party algorithms as well as our own if we rely on them for any activities, need to create roles to test the algorithms, importance of awareness of both bias in the data - e.g. through subsets selected and bias in the algorithms - from the code/human trainers and need for much better education of executives on the risks and opportunities of AI.
A few bilateral meetings and an incredibly interesting dinner with a retired US 4star general as speaker...I’ve learnt lots of new things about ‘hyper war’, how AI and bioengineering could make things a lot better or a lot worse and the importance for business to step up in driving the ethical debate...probably a good note to end on!
Martin's pre-conference blog can be found here.
Her round up of day 2 can be found here.
Her final round up can be found here.
