Presented By
O’Reilly + Cloudera
Make Data Work
29 April–2 May 2019
London, UK

Why is it so hard to do AI for good?

Duncan Ross (Times Higher Education), giselle cory (DataKind UK)
17:2518:05 Wednesday, 1 May 2019
Law and Ethics, Strata Business Summit
Location: Capital Suite 12
Secondary topics:  Ethics
Average rating: ****.
(4.50, 4 ratings)

Who is this presentation for?

  • Data scientists interested in data for good

Level

Beginner

What you'll learn

  • Understand the challenges of data for good and how to address issues such as bias in the most challenging of all sectors: the social sector

Description

How do you identify social challenges that are appropriate for data science? How can you avoid the biggest pitfalls, and how can you ensure solutions are sustainable?

The reality is that it’s easy to create biased or racist models and spin up rough-and-ready prototypes. It’s hard to build innovative solutions that are sustainable and have positive social impact. Traditional business models often don’t work when your customer base is a vulnerable community. Leapfrogging institutions doesn’t work as there needs to be genuine partnerships with those working on the social problems.

Duncan Ross and Giselle Cory explore how DataKind and its volunteers have tackled these problems (and yes, they will admit to their mistakes). Join in to gain insights into how to identify the right AI or data for good cases and how this can act as a framework for avoiding the same problems across industry.

Photo of Duncan Ross

Duncan Ross

Times Higher Education

Duncan Ross is chief data officer at Times Higher Education. Duncan has been a data miner since the mid-1990s. Previously at Teradata, Duncan created analytical solutions across a number of industries, including warranty and root cause analysis in manufacturing and social network analysis in telecommunications. In his spare time, Duncan has been a city councilor, chair of a national charity, founder of an award-winning farmers market, and one of the founding directors of the Institute of Data Miners. More recently, he cofounded DataKind UK and regularly speaks on data science and social good.

Photo of giselle cory

giselle cory

DataKind UK

Giselle Cory is the executive director of DataKind UK, an organization that helps the social sector use data science by bringing together social change organizations and pro bono data scientists. Previously, Giselle worked in the UK government (including in the Prime Minister’s Strategy Unit), and for national charities and think tanks (including the Resolution Foundation), using data to better inform public policy decisions. She believes that smart, responsible data collection and use can help the social sector tackle some of the UK’s biggest challenges. Giselle holds an BSc in maths and physics, a MSc in computational journalism, and diplomas in economics and manned spaceflight.