Bias in AI – what next?
Women, ethnic minorities, the poor and disabled are just some of the groups discriminated against because of bias in machine learning. Can the bias in AI ever be removed?
Our own bias is becoming engrained in computer code. There is a huge amount of evidence showing that human bias and ignorance is encoded into our digitally driven world. The impact of this is unsurprisingly impacting the most vulnerable communities the hardest – decisions on health care, employment and even police surveillance are now being made very often by machines. But can anything be done to stop this bias from getting any worse and can the current bias be removed? As part the WebSci 2021 conference Digital Planet looks at what can be done by public bodies and the private sector to improve AI ethics.
Joining us are Professor Lucy Hooberman, Professor Ricardo Baeza-Yates, Dr. Rumman Chowdhury and Dr. Margaret Mitchell.
(Image: Getty Images)
The programme is presenter by Gareth Mitchell with expert commentary from Bill Thompson.
Studio Manager: Bill Thompson
Producer: Ania Lichtarowicz
Last on
Clip
![]()
How exclusion bias affects important data sets
Duration: 03:30
Broadcasts
- Tue 22 Jun 202119:32GMTBBC World Service
- Wed 23 Jun 202103:32GMTBBC World Service Australasia, South Asia & East Asia only
- Wed 23 Jun 202104:32GMTBBC World Service Americas and the Caribbean
- Wed 23 Jun 202108:32GMTBBC World Service
- Wed 23 Jun 202112:32GMTBBC World Service except East and Southern Africa, East Asia, South Asia & West and Central Africa
- Sat 26 Jun 202116:32GMTBBC World Service News Internet
Podcast
![]()
Digital Planet
Technological and digital news from around the world.
