Judgment Approved by the court for handing down
R (Bridges) v CCSWP and SSHD
35. Before it is possible to draw conclusions on the existence of
gender bias, an extensive study needs to be conducted where
match scores are thoroughly analysed for both males and females,
regardless of whether they generate alerts or not. Mr. Edgell does
not carry out this study; he considers only alert statistics.”
(“Lambs” is a label used by the software providers to describe faces that have
a number of common generic features such that more frequent matches are
generated by the facial recognition software.)
156.
Thus, SWP may now, in light of the investigation undertaken to date by Mr.
Edgell, wish to consider whether further investigation should be done into
whether the NeoFace Watch software may produce discriminatory impacts.
When deciding whether or not this is necessary it will be appropriate for SWP
to take account that whenever AFR Locate is used there is an important
failsafe: no step is taken against any member of the public unless an officer
(the systems operator) has reviewed the potential match generated by the
software and reached his own opinion that there is a match between the
member of the public and the watchlist face.
157.
Yet this possibility of future action does not make good the argument that to
date, SWP has failed to comply with the duty under section 149(1) of the
Equality Act 2010. Our conclusion is that SWP did have the due regard
required when in April 2017 it commenced the trial of AFR Locate. At that
time, there was no specific reason why it ought to have been assumed it was
possible that the NeoFace Watch software produced more or less reliable
results depending on whether the face was male or female, or white or
minority ethnic. As we have explained, even now there is no particular reason
to make any such assumption. We note that although Dr Jain states that “bias
has been found to be a feature of common AFR systems” he does not provide
an opinion on whether, or the extent to which, such bias can be addressed by
the fail-safe, such as ensuring that a human operator checks whether there is in
fact a match.
158.
In our view, the April 2017 Equality Impact Assessment document
demonstrates that due regard was had by SWP to the section 149(1) criteria.
The Claimant’s contention that SWP did not go far enough in that it did not
seek to equip itself with information on possible or potential disparate impacts,
based on the information reasonably available at that time, is mere speculation.
In any event, as matters had developed in the course of the trial since April
2017, it is apparent from Mr. Edgell’s evidence that SWP continues to review
events against the section 149(1) criteria. This is the approach required by the
public-sector equality duty in the context of a trial process. For these reasons,
the claim made by reference to section 149(1) of the Equality Act 2010 fails.
H.
CONCLUSION
159.
For the reasons set out above, the Claimant’s claim for judicial review is
dismissed on all grounds. We are satisfied both that the current legal regime is