An update on The OR Society’s response to the challenges and opportunities of artificial intelligence


It has been a while since I last provided an overview of how the ORS has been positioning itself to respond to the opportunities and challenges that arise from the rapid evolution in artificial intelligence (AI). Looking back, my last Leader in Inside OR on the topic was in autumn 2019. Much, of course, has happened since then, not least the ongoing issue of COVID-19 (which, you may be pleased to know, I don’t plan to say much about here!)

I’ll start with a condensed version of the backstory over the last couple of years and then say a little about the work that we’re currently undertaking and how anyone interested can get involved.

Alan Robinson.png

In late 2018 – long before COVID-19 (oops, mentioned it again!) – at the behest of the Analytics Development Group (ADG), the ORS Board agreed to set up a small Task and Finish (T&F) Group to consider two key issues. First, we wanted to come up with a simple way of explaining the interaction and overlap between OR and AI in all its guises to both those within the OR community and more widely; and, most particularly, the relationship between OR and machine learning (ML – and related techniques) as that is where some of the most apparent overlaps in approaches potentially sit. And, secondly, we wanted to consider how ORS – and the broader OR community – might best respond both to the opportunities presented and also, where required, to any challenges.

The core membership of the T&F group was deliberately kept small whilst ensuring that it included representative viewpoints from both academia and practice (private sector and government) as well as ORS staff. We conducted most of our work virtually using a combination of telecoms and a shared website to facilitate group working and document sharing, but that was mainly for diary management purposes rather than because of any travel restrictions. With hindsight, this proved most fortuitous! However, we did have one very useful full faceto-face meeting and, importantly, also undertook a number of wider consultations. These allowed us to gather wider inputs and, particularly later in the group’s work, road-test our emerging ideas.

We included a session with ORS General Council in Spring 2019, a survey of ORS membership that summer, discussions with HORAF (Heads of OR and Analytics Forum), and a workshop at OR61 as part of Making an Impact activities that covered the group’s key findings; as well as regular updates for – and discussions with – ADG and ORS Board. The work of the T&F group was also shared with INFORMS, and by way of reciprocation, we received a paper that they were working on at the time that covered very similar ground to the T&F deliberations.  We continue to have a productive exchange with INFORMS around AI and ML to this date.

Ultimately, the T&F group reported to the board in Dec 2019; its key recommendations, which I will now outline, were accepted. The first main recommendation was that ADG should formally embrace AI as part of its remit, revisiting its Terms of Reference and membership to ensure that it could discharge that broader remit. This extension to ADG’s core role – which arose originally when the ORS (and the wider OR community) needed to address similar questions around the emergence of data science and analytics – allows ADG to provide the focal point for ORS discussion around AI and ML.

However, the impact of AI and ML is pervasive, so the second recommendation area concerned activities that needed to be picked up elsewhere in ORS. For example, the need for a revised training offering is being taken forward by the Education Committee to make sure that AI/ML related training is available to members. The Events Committee has been taking forward both new events in the AI/ML field (such as Validate AI) as well as adjusting the content of extant events (such as more comprehensive AI/ML coverage at the annual conference as well as at the Analytics Summit), and Publications continues to ensure that the range of ORS journals covers relevant needs (such as via the Journal of Business Analytics).

And, lastly, ORS Board agreed to a couple of wider recommendations that emerged from the work.

One area worth mentioning here is the need for wider consideration of ethics and biases: and the provision of best practice guidance in those areas. This is currently being led by ADG – and is likely to cover more than AI/ML – so watch out for further details as that work evolves.

So, in closing, where are we now? ADG has recently undertaken a short review on behalf of the ORS Board to check on progress.

I’m pleased to say that this showed that the relevant activities are now embedded across ORS; however, as ever, there is more to be done, not least because of the pace at which AI/ ML continues to evolve. We will also be looking to share more detail and updates via the ORS website over the coming months.