Ms Doireann Ansbro:

It is important that consideration should be given to future technologies, as well as to facial recognition technology, given the potential impact on minority communities and the potential for discrimination. We have already seen through research in other jurisdictions that using facial recognition technology has a negative and disproportionate impact on people of colour and on women. This is because of misidentification, as well as because of biases that are built into, as the Senator said, the algorithms in that the unconscious biases of the developers are often reflected within the development of this software.

In terms of how to protect people, it is important to note that the EU in the proposed regulation on artificial intelligence has suggested that facial recognition technology should never be used in a public place. We need to be cognisant of that. I would highlight that this is contrary to what is in An Garda Síochána modernisation plan 2016-2021, which states that it plans to use facial recognition technology to track suspects from CCTV, as well as to use technologies such as face-in-the-crowd and shape-in-the-crowd biometrics. We would be highly concerned about any use or potential use of facial recognition technology. We need to specify that that will not be used in the Irish context.

It is also important to note that the current Department of Justice and An Garda Síochána code of practice states that automatic facial recognition technologies are prohibited. In our opposition to body-worn cameras, we have also noted concerns in other jurisdictions, where they have been more used in communities that are already over-policed. This is having an impact on trust and on community policing between the people who are policed and the police themselves. It also potentially impacts on garda discretion itself, where a garda is able to use his or her discretion on whether or not something is pursued. There are a lot of ramifications and this is why we are calling for more research and for a pilot project before there is a general roll-out. There are a lot of ramifications that we need to consider before such software is rolled out.

One other point is about this idea that potentially new software can be included in this legislation. I understand the rationale is that safeguards should be applied to any future technology. However, we need to be cognisant that the rapid development of technology and software at the moment poses new and unforeseen risks to rights. That is why we say that future technology should be assessed individually, on a case-by-case basis, to make sure that the impact on rights is not disproportionate. A dangerous part of this legislation is that new software is potentially envisioned to be covered by it.