While this speaks to unconscious bias that is seen in both the UK and the US populations, mechanisms to discourage bias, both implicit and explicit, in DNA collection policies would help redress the imbalance.
Professor Denise Syndercombe Court
28 April 2021
The need for a protocol to prevent racial bias in DNA databasing
Professor Denise Syndercombe Court highlights the potential for racial bias in the way DNA profiles are collected and stored.
The proportion of non-white people represented in DNA databases in both the United Kingdom (UK) and United States (US) significantly outweighs that of the population in general. This can largely be attributed to the way criminal DNA is collected, as collection protocols do not always protect against bias – whether unconscious or conscious.
Denise Syndercombe Court, Professor of Forensic Genetics at King’s, asserts that the over-representation of black people on the UK National DNA database can only increase the feeling that they are being unfairly criminalised, affecting a high proportion of families.
Similarly, in the US, individuals classified as Black are between two and three times more likely than white people to have DNA collected.
Law enforcement personnel have been known to cast wide dragnets to obtain DNA profiles when investigating crimes, to increase their chances of finding perpetrators. However, this practice means that the DNA profiles of large numbers of people, often non-white people, are collected and stored on local or national databases.
New DNA typing technologies have enabled rapid on-site DNA profiling, currently being proposed for use in US custody suites. But it is the ungoverned and misguided opportunity to collect and retain DNA profiles for local use in unregulated databases, particularly seen in the US, that needs to be addressed.
Pointing to a computational solution proposed by Blindenbach and colleagues in Nature Computational Science, Professor Syndercombe Court affirms that a system in which DNA profiles are not revealed (and so cannot be 'known’) but are instantly compared to a national database could help to solve the problem.
This system - by providing a simple match/no-match output for law enforcement - would allow for useful and timely custody decisions to be made while avoiding back-door solutions that can be racially biased.
Read Professor Syndercombe Court’s full comments: Protecting against racial bias in DNA databasing, published in Nature Computational Science this week.