The UK National Data Strategy takes a v optimistic line on algorithmic and data bias in S. 2.5: “We can harness data as part of struggles to tackle bias and exclusion. Data can be used to hold a mirror up to society … Biases will need to be addressed”, and asks for input (1/n)

The consultation question is “please provide any comments about the potential impact the proposals outlined in this consultation may have on individuals with a protected characteristic”. I am oversimplifying, but the proposals are moreorless … (2/n)

Strengthening incentives to share data across the public sector, with “appopriately safeguarded, joined-up, and interoperable data infra” and also proposals to enable better data sharing between biz, civil society and gov. (3/n)

Frameworks, governance and ethics are all mentioned, but not in detail, and there is a hat tip to making the CDEI a statutory body. I’m going to be optimistic and say this is an opportunity to influence and outline some good practice, and assert some red lines (4/n)

There is a great deal of scholarship and journalism that demonstrates why this might not just be problematic, but also unjust and exacerbate bias (for instance, this great work by @Abebab https://venturebeat.com/2020/07/01/mit-takes-down-80-million-tiny-images-data-set-due-to-racist-and-offensive-content/) (5/n)

I am sure some of the more academically oriented #digitalethics orgs will address this, but it also seems like an impt issue for civil society orgs, who may not have data expertise, but may be mitigating the impact of biased data or working against structural bias/injustice (6/n)