By March 2020, the COVID-19 pandemic had begun to create huge difficulties in the daily lives of most people the world over. Working environments began to see heavy restrictions, schools and universities closed for months and the concept of working from home en masse became the norm, presenting all its advantages and disadvantages. Whilst most people will remember the pandemic as a very difficult period in their lives, some remarkable positives emerged. Every crisis comes with opportunities and COVID was no different, opening doors that, to a certain extent, had already been ajar. In the modern age, most areas of academic research are highly internationalised and therefore were hit tremendously by restrictions on travel and one-to-one communication. Conferences and workshops were cancelled, young researchers found themselves with significantly less personal attention from their supervisors and informal discussions at seminar talks came to an end, literally overnight. In a very short time, many new initiatives were developed to explore different ways to maintain research communication and the results went much further than anyone could have previously imagined.

During the very first lockdown weekend, sitting at home on a Sunday evening towards the end of March 2020, we were polishing a joint article and discussion drifted to the new-found virtues of the webtool Zoom, the platform through which we were communicating. Like many of our colleagues around the world, Zoom suddenly came to prominence at the moment that there was an urgent need to transform teaching from the classroom to the virtual world. The sudden global familiarity with Zoom and online mathematical engagement presented a possible opportunity. The idea of virtual seminars and conferences has been circulating for some time already among the mathematical community. The principal motivation for them has been to promote inclusion together with the need for the mathematical community to reduce its carbon footprint. Discussions towards making changes in this direction had previously always lost momentum for a variety of reasons, but now the time was right to move on.

After an animated discussion about what setting up a global seminar would entail, we decided there was no point to hesitate and we should commit ourselves to host a brand-new seminar series by the end of the week. Although it seemed foolishly ambitious, we conjectured that the simplicity of what needed doing made our target rather easy to realise. Indeed, just four days later, Nathanael Berestyki from Vienna gave the very first One World Probability Seminar (OWPS) in front of a global crowd of an audience of 450 probability researchers, spanning the West Coast of the USA to Japan.

The set-up was remarkably simple, requiring just five easy steps. We guestimated how many people might turn up and paid a little bit extra to have a larger Zoom licence (initially the Zoom Pro 500); we made sure that we understood how to control large audiences on Zoom;  we wrote to a series of potential seminar speakers, ensuring that we had at least 2-3 weeks of seminars lined up to announce at the first session; we set up a rudimentary webpage in a matter of minutes; and finally we sent a single email to around 200 recipients around the world, asking them to spread the word that in two-days-time the first seminar would be held. It was the latter that surprised us the most. The idea clearly resonated and the email spread like wildfire. We knew this because, within hours of sending the initial email, we had begun to receive copies of the email via other network forums. It was seeing this volume of email traffic that concerned us enough to upgrade the Zoom licence at the last minute before the initial broadcast to a Zoom Pro 1000.

Since the very first broadcast, the OWPS setup has remained mostly unchanged. It consists of two 45-minute talks with a quick coffee break at home followed by the opportunity to discuss with the speakers. Almost from the get-go, it was clear that finding a perfect time to make OWPS a true ‘world’ seminar instead of a ‘half-world’ seminar would be difficult. The initial OWPS timetable was programmed in to be late afternoon central European time. This temporal positioning meant that Latin America, the East Coast of North America, the whole of Europe, India, Russia and as far East as Beijing could comfortably attend. Some early risers on the West Coast of North America were also in regular attendance. Talks were recorded and put on YouTube to give access to everyone who could not attend live, or wanted to see the talks again. The aims of OWPS were publicised among participants as advancing a more environment-friendly research culture, being as inclusive as possible, and providing access to high-level research seminars to everyone everywhere.

We believe that we achieved much more than we initially hoped for. Along with many other virtual seminars and conferences that were kick started around the same time, OWPS demonstrated that there are multiple ways in which virtual platforms can be used to this end. The principle that it is not necessary to fly speakers to a venue in order for them to give a successful seminar was also clearly established.

The concept was very quickly adopted by other mathematical communities. Within a week of our first seminar, the One World PDE Seminar was up and running, pulling in even bigger crowds than OWPS, followed quickly by One World Mathematical Game Theory Seminar and the Waves in One World Seminar. To date, there are no less than 22 seminars in the ‘One World’ series, covering a huge range of mathematical specialities, but also one in Cognitive Psychology.

Whilst the experience has demonstrated some overwhelming positives, there are also some important drawbacks of virtual seminars that are worth noting. Often the two come hand-in-hand. A good example of this pertains to the ability to invite leading mathematicians to give talks; such invites are easy to extend and equally easy to accept as they cost little more than an hour of contact time over the internet. That said, the temptation to fill seminar series with highly prestigious speakers does carry a downside. Although offering an opportunity, this does distract from the need of young researchers to present and discuss their research.

This particular issue was specifically addressed by the IMS-Bernoulli One World Symposium, which took place in August 2020, which was set up, again at short notice, as an alternative to the physical meeting that would have taken place at the same time had the pandemic not hit. Noting the diminished opportunity for young researchers to speak at a variety of COVID-hit forums, the aim of the symposium was to open the stage specifically to them. With some very light-touch vetting, participants of the symposium were invited to submit a pre-recorded 10-minute video followed by the opportunity for inverted classroom-style live discussions during the symposium week, making full use of the 24hrs in the day across the different time zones. This idea was somewhat of a gamble as it was unclear what the uptake would be and whether early career researchers, in particular, would be willing to engage.

As it happened, the gamble paid off. The One World Symposium ended up with almost 600 pre-recorded talks, with the vast majority of them submitted by young researchers. The discussion sessions (which were spread across 23 different themes) were also heavily populated with numbers at the more popular discussion sessions regularly topping out above 50. Despite the recorded talks being restricted to 10 minutes, the speakers all used the opportunity to present their work in a very efficient way; making them wonderful viewing, offering seemingly the same amount of content as many 45-minute live seminars offer.

It appears that such formats could, in principle, replace large, general-audience research conferences. Indeed, it is arguable that the aim of gaining an overview of developments in different fields is much easier fulfilled by viewing a variety of short pre-recorded talks and attending the accompanying discussion sessions than by attending the few sessions one can attend during a normal conference. The administrative burden in setting things up also proved to be dramatically much work less as well.

The beginning of the pandemic saw just a few, albeit heavily attended, virtual seminar series. These paved the way for the now increasing number of more specialist and (temporally if not regionally or linguistically) localised seminars and virtual conferences.  It seems difficult to predict the future of virtual activities as the pandemic progresses, as well as what will happen once the constraints of the virus are over. Most certainly there are advantages of virtual seminars that go well beyond environmental factors, which, we believe, will mean that virtual forums will outlive the pandemic. Being (geographically) inclusive and family-friendly are major factors in this respect. Researchers without a travel budget can participate in events that were completely out of reach before the pandemic. Those with caring responsibilities can participate in international events without leaving their families for a substantial amount of time. Nonetheless, it must be said that the pleasure and benefits of spending time with colleagues, both at the board or socially, can never be entirely replaced by the virtual forum.