Artificial Intelligence, how banks need to keep ahead of emerging technology. Communicating future media and medical practices as a part of Science Communication as Live Performances.
The Future of Ethical AI Frenzy Panel:
The third AI Frenzy at Eagle Lab Salford. The event focuses on the future of Ethical AI — a topic that’s increasingly being discussed and debated among AI experts. Featuring: Dr David Rose, Emma Erskine-Fox, Tom Begely, Niels Pedersen, James Medd and panel host Paul Wilshaw.
AI Frenzy — The Future of Ethical AI — Live Broadcast:
Broadcast on YouTube Live from The Landing, MediaCityUK, with presentations from speakers Dr David Rose, Emma Erskine-Fox, Tom Begely and Niels Pedersen. Wrapping up with a panel discussion chaired by Paul Wilshaw. The panel discussion is broadcast in audio only with the video failing to stream.
🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖
In this performance, AI Frenzy — The Future of Ethical AI. I host a panel discussion about artificial intelligence (AI). The purpose of which is to explore the different aspect of ethical AI from the perspectives of myself and the panel. The panel included Emma Erskine-Fox is a technology and data protection lawyer from TLT; Niels Pedersen, Senior Lecturer in MSc Financial Technology at MMU; James Medd, artist, educator and technician at The
Landing; Tom Begely Data Scientist for ASI Data Science and Dr David Rose Lecturer in Geography (UEA) currently working with Defra.
Through this video, I aim to engage the audience, of around 90, focusing on changing attitudes and emotion rather than a cognitive practice of comprehension or application, often the emphasis in science education (Lesen, Amy E., et al., 2016). I never aimed to make the audience understand AI or the numerous future ethical positions. I did want to engage and provoke a change in attitude towards AI from users of the technology to those
creating it.
The purpose of the series of AI Frenzies is to promote Barclays as a future thinking and doing bank rather than the traditional “bricks and mortar” branch network. Since the financial crisis of 2008, where a large number of financial institutions collapsed or were bailed out by governments (Erkens, Hung and Matos, 2012), with an 80% drop in perceived lack of trust in the financial markets (Sapienza and Zingales, 2012). Since the crisis, banks have had to change their representation significantly to the general public. During this time technology has changed dramatically; how people interact and expect instant access to their banking services. Financial service providers have faced three main challenges when digitising their service for the future economy; connection: how to unite users, merchants, and service partners in a low-cost, fast, and intelligent process; risk: how to manage aspects of financial risk and emerging hackable technology platforms; and trust: how to grant equal opportunity for all to be trusted, and trustworthy, in the digital
space (Qi and Xiao, 2018). While AI has been around for a long time its integration into financial services has been slow. AI enables banks to build intelligent systems that better serve customers and business partners through algorithms that determine, for example, the outcome of a loan application. Numerous decisions precede this and often judgement by the individual assessing the risk of the application. A single development error can be
catastrophic, as in the case of Knight Capital, who lost $400 million, due to a bug in their programme (Silicon Valley Innovation Center, 2019). The example of Knight Capital is accountable through data, we cannot reliably track biases in the information we serve to machines reflects our history of unequal society and unconscious prejudices, we are, in effect, asking the program to learn our own biases and flaws (Buranyi, 2019). With this
history and potential future, the AI Frenzy event intends to educate and entertain the audience without any collaboration on AI products. I emphasis entertain as that is key to creating a lasting engagement with the audience and not to lose my connection with them to their ever notifying devices.
With this background in mind, I organised the event, in particular, the panel discussion, to provoke thought and consideration when producing an ethical AI product — in all parts of the production and not just development. Asking the audience to think about their own unconscious biases if they were to create or work within AI. The panel consists mainly of white bourgeoisie men, limiting inclusiveness of thought leadership. I reference this during the performance and suggest the audience to consider including a comprehensive
demographic when defining the underlying algorithms that could determine a severe consequence for the consumer.
The panel discussion takes its narrative cues from a predetermined outline of subjects consisting of robotic workforces, smart speakers, financial data and agricultural technology (agri-tech). As well as being influenced by the precursor lightening talks (talks of ten minutes or less) to draw on the attendees previous experience at the event. I aimed for the content to be expository or argumentative communication, or with other samples of explanations, such as descriptive, deductive, or analytical. Science communication narratives are usually compared with the logical-scientific communication they are underlying. Logical-scientific communication strives to provide abstract truths that are proven across specific situations (Dahlstrom, 2014). The panel references are broad generalisations of their experiences to communicate their inductive reasoning. These are in no way absolute facts about ethical AI even the speakers acknowledge this and attendees are encouraged to draw their conclusions on the subject matter.
How is effective was this particular panel discussion? In comparison to one of the most significant AI events is the World Summit AI held in Amsterdam: https://youtu.be/Ff3opeyvoSk — probably not as effective. The World Summit AI evaluates their conference through attendees, speakers, global reach and subject areas, or tracks. With 46 sponsors for the 2018 summit and 6000 attendees, that is around 130 attendees per sponsor (World
Summit AI Amsterdam, 2018). I acknowledge that this is not a final measurement of success, but it does provide an approximate benchmark for attendance. Perhaps then an audience of 90 in a non-capital city is not bad going?
Barclays main aim with its AI Frenzy events is to promote trust in emerging technology. AI and the fourth industrial revolution are critical to the established financial service providers staying ahead of their challenger banks, such as Monzo, Starling Bank and Revolute. Barclays has to retain not only their customer base but their technology teams to deliver the 24/7 always-on customer experiences that are readily available in other sectors. Less financial pressures are facing AI Frenzy than, say a small start-up. There is a balancing of cost recovery versus scholarly impact, customer engagement, employee recruitment and retainment can mean walking a tightrope of funding and messaging (Radko, 2015). By no means am I an accomplished funambulist, such as Phillippe Petit, I do try and balance subjectivity with the overarching aim to the event. I hope the balance was suitable and enjoyed by the attendees as this tweet from @EmmaFoxLaw suggests:
“A pleasure to speak at the #barclaysaifrenzy #ethicalAI event this evening. Some great discussions with this brilliant panel, expertly chaired by @paulwilshaw” (Erskine-Fox, 2018).
[image] Available at: https://twitter.com/EmmaFoxLaw/status/1073350203984486400.
🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖
Along the way, I considered differing subjects and platforms to use from creating a podcast to broadcasting live. There are numerous things to consider. Would my passion be observed, acknowledged and understood by an unseen public in the realms of social media? What would be of interest to people who did not know me? Could I grab their attention-hungry gaze for the entire duration of my performance? Ideally, I want to create the perception of “FoMO” (Fear of Missing Out) — the anxiety that an exciting or interesting event may be happening somewhere else on social media (Hogan, 2015) and not to be a victim of FoMO. I decided that using YouTube was the best option as this can be done directly from a smartphone, while other platforms have the same option, such as Periscope (by Twitter) Facebook Live and Instagram, I found YouTube feature-rich and provide a level of ambiguity — I don’t have many followers or use the service regularly. I tend to separate my activity from platform to platform, and my external persona on each differs slightly for its intended use. To me, YouTube is a “blank canvas” unexplored, except for the occasional reference video and often seen in the form of embedded code on other sites.
🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖
I also created an audio podcast and video podcast with fellow student Damon Reid to discuss science fiction and as the late sci-fi author Thomas Disch called it “creative visualisation” — in a looping process where science fiction writers imagine the substance of our futures and the generation raised on those visions construct them into reality. Space X, smartphones and facial recognition all evolved in this manner (Merchant, 2018). With inspiration from podcasts such as Science Cafe by BBC Wales (Available at: https://www.bbc.co.uk/programmes/m0001xy6) and Brainwaves by BBC Scotland, I wanted to experiment with a more unstructured approach than that of the AI Frenzy using minimal equipment and no set-up — as a result, Techs on Trams was born. Using just a smartphone to film a tram journey with myself and Damon allowed us to capture the episode and broadcast using YouTube Live over a 4G network. Initially, I recorded Damon holding the phone, in landscape mode, while he talked about virtual reality (VR), augmented reality (AR) and mixed reality (MR) on the moving tram. This method worked okay but didn’t portray Damons’ natural expertise on his subject of choice. I felt the rapidly crowding tram, unsuspecting and reluctant audience did not enhance the situation. I also felt conscious behind the camera that none of the audience had explicitly agreed to be part of this performance. Damon and I swapped roles to film my segment on science fiction writers and the predictions of their future and our present. I felt anxious due to the accidental audience of commuters and rushed my performance to the camera. After filming and live broadcasting both performances we agreed to perform on screen together. This shared experience and past performances put us both more at ease and relaxed in front of the camera. We both felt that this version of the video podcast was the better performance. Although more technically challenging — I was holding the phone at arm’s length while trying to keep two people of dramatically different heights in the frame on a moving tram, listening for audio spikes (notably noisy commuters) while talking is unquestionably a talent.
Techs on trams 👩💻👨💻💬🚉with Paul Wilshaw and Damon Reid:
Take a tour into the present and future of technology, with Damon Reid and Paul Wilshaw. Damon and Paul look into the possibilities of VR, what it has been used for now and what it could be used for — all on a tram.
Techs on trams 👩💻👨💻💬🚉, part one:
Damon Reid takes you on a journey into Virtual Reality (VR) and Augmented Reality (AR).
Techs on trams 👩💻👨💻💬🚉, part two:
Paul Wilshaw takes you on a journey into the present and the future of technology. The original Blade Runner (1982) film was set in 2019. So what predictions did the film get right and what wasn’t predicted?
Techs on trams 👩💻👨💻💬🚉 playlist:
https://www.youtube.com/playlist?list=PLjCEbS9K1Va8OB-SWQOzN_05LpXTdYCcZ
Techs on trams 👩💻👨💻💬🚉 podcast:
https://www.dropbox.com/s/wk5lu400csonb1h/Techs%20On%20Trams%20-%20Final%20Mix.wav?dl=0
🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖
I also helped Karina Silabriede, year one MSc Science Communication student, to make a live video broadcast on the subject of cardiac arrests in the hospital and her vast experience as a healthcare professional. We decided to film it live, after a few off-camera rehearsals, in the setting of the cinema room at The Landing, also used by The Voice UK, we commenced filming, again using a smartphone and YouTube live. I took the role of host and Karina the subject expert — even demonstrating resuscitation techniques.
Science communication — Cardiac Arrest, in the hospital 🏥 — take one:
https://youtu.be/Ul0tkeyIv1o
Science communication — Cardiac Arrest, in the hospital 🏥 — take two:
https://youtu.be/FCMbmwmjT0k
🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖
In conclusion, AI Frenzy — The Future of Ethical AI was the more comfortable performance. I had previously been on the panel of AI Frenzy — Women in AI (https://vimeo.com/288512018) and presenting on the topic of AI, ethics and citizen science (https://vimeo.com/290239213/76daee38e9). I knew what to expect — the kind of audience attending, the event composition and the shared experience of the panel to bolster my performance. Similarly, I knew that my YouTube Live broadcast using just my smartphone gaffer taped to a table was an experiment, an addendum to that of being filmed by a high-end camera operator. I had time before to prepare the room, brief the panel, adjust the lighting and arrange the seating — all carefully orchestrated to engage the audience and viewer to create the aspired FoMO experience. The evaluation of the performance produced satisfying engagement with the audience who attended than those who watched the video — if measured by reaction and interaction. The video hosted on Vimeo currently has three views, 36 impressions and only one viewer watching to the end. Not accounting traffic from different sources, such as websites, blogs, television and exhibitions. Nor does it measure topic engagement and if anyone has consciously or subconsciously changed their behaviour due to this performance.
Initially, Techs on Trams was intended to be an audio podcast. However, upon research, the benefits of video podcasts are evident. Robin H. Kay at the University of Ontario confirms this by reviewing a number of studies where almost half observed cognitive reactions toward video podcasts and over 85% of the conclusions were positive. Kay notes that one study proposed the visual nature of video podcasts helped to explain the topic and increase understanding (Kay, 2012).
A study by Ullrich Wagner, Lisa Galli, Björn H. Schott, Andrew Wold, Job van der Schalk, Antony S. R. Manstead, Klaus Scherer and Henrik Walter applies subjective ratings and functional magnetic resonance imaging (fMRI) data from participants who viewed positive, neutral or negative images from inside an MRI scanner. A friend either watched (shared emotional experience) or did not watch (unshared emotional experience) the same images simultaneously. This study illustrates how emotional stimuli with a friend buffers the impact of negative stimuli and enhances the effects of positive stimuli (Wagner et al., 2014).
As with the other performance examples, the live video with Karina was done with the emotional buffer of a friend and faired better for it. One of the most significant qualities of human beings is their social nature. People have a strong urge to form social bonds and to share their experiences with them (Wagner et al., 2014). All the performances are experienced with others, either immediately during their creation and making; or upon uploading to a social media platform. I do not believe that the platform is of relevance — it is a preference and tool to form a social bond and communicate messages; monitoring effectiveness over time, experimented with and continuously added to.
🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖 🤖 💬 🤖
References
BBC Wales (2019). What’s in store for 2019?. [podcast] Science Cafe. Available at: https://www.bbc.co.uk/programmes/m0001xy6 [Accessed 8 Jan. 2019].
Buranyi, S. (2019). Rise of the racist robots – how AI is learning all our worst impulses. [online] the Guardian. Available at: https://www.theguardian.com/inequality/2017/aug/08/rise-of-the-racist-robots-how-ai-is-learning-all-our-worst-impulses [Accessed 13 Jan. 2019].
Dahlstrom, M. (2014). Using narratives and storytelling to communicate science with nonexpert audiences. Proceedings of the National Academy of Sciences, 111(Supplement_4), pp.13614–13620.
Erkens, D., Hung, M. and Matos, P. (2012). Corporate governance in the 2007–2008 financial crisis: Evidence from financial institutions worldwide. Journal of Corporate Finance, 18(2), pp.389–411.
Erskine-Fox, E. (2018). Twitter. [online] Twitter.com. Available at: https://twitter.com/EmmaFoxLaw/status/1073350203984486400 [Accessed 13 Jan. 2019].
Hogan, M. (2015). Facebook and the ‘Fear of Missing Out’ (FoMO). [online] Psychology Today. Available at: https://www.psychologytoday.com/us/blog/in-one-lifespan/201510/facebook-and-the-fear-missing-out-fomo [Accessed 13 Jan. 2019].
Kay, R. (2012). Exploring the use of video podcasts in education: A comprehensive review of the literature. Computers in Human Behavior, 28(3), pp.820–831.
Lesen, A., Rogan, A. and Blum, M. (2016). Science Communication Through Art: Objectives, Challenges, and Outcomes. Trends in Ecology & Evolution, 31(9), pp.657–660.
McGillion, C. and McKinnon, M. (2014). Participatory Theater as a Science Communication Tool in Timor Leste. Science Communication, 36(4), pp.502–510.
Merchant, B. (2018). Nike and Boeing Are Paying Sci-Fi Writers to Predict Their Futures. [online] Medium. Available at: https://medium.com/s/thenewnew/nike-and-boeing-are-paying-sci-fi-writers-to-predict-their-futures-fdc4b6165fa4 [Accessed 13 Jan. 2019].
Qi, Y. and Xiao, J. (2018). Fintech. Communications of the ACM, 61(11), pp.65–69.
Radko, T. (2015). Measuring Success: The Value of Our Work Can’t Always Be Captured in a Spreadsheet. Choice Reviews Online, 52(09), pp.1457–1462.
Sapienza, P. and Zingales, L. (2012). A Trust Crisis. International Review of Finance, 12(2), pp.123–131.
Silicon Valley Innovation Center. (2019). AI- an Opportunity or Threat for the Finance Sector?. [online] Available at: https://siliconvalley.center/blog/how-artificial-intelligence-impacts-the-finance-industry-opportunity-or-threat/ [Accessed 13 Jan. 2019].
Twitter (2018). Screenshot of Twitter @EmmaFoxLaw on reflection of AI Frenzy — The Future of Ethical AI. [image] Available at: https://twitter.com/EmmaFoxLaw/status/1073350203984486400 [Accessed 11 Jan. 2019].
Vimeo (2019). Barclays The Future of Ethical AI Frenzy Panel Part 3. [image] Available at: https://vimeo.com/manage/307888942/stats#video [Accessed 13 Jan. 2019].
Wagner, U., Galli, L., Schott, B., Wold, A., van der Schalk, J., Manstead, A., Scherer, K. and Walter, H. (2014). Beautiful friendship: Social sharing of emotions improves subjective feelings and activates the neural reward circuitry. Social Cognitive and Affective Neuroscience, 10(6), pp.801–808.
World Summit AI Amsterdam. (2018). World Summit AI | The World’s Leading AI Summit | Oct 2019 | Amsterdam. [online] Available at: https://worldsummit.ai [Accessed 13 Jan. 2019].
YouTube (2019). Screenshot of Techs on trams 👩💻👨💻💬🚉with Paul Wilshaw and Damon Reid.. [image] Available at: https://youtu.be/CGGj9eLtJys?t=11 [Accessed 13 Jan. 2019].
YouTube (2019). Screenshot of Science communication — Cardiac Arrest, in the hospital 🏥 — take two.. [image] Available at: https://youtu.be/FCMbmwmjT0k?t=1063 [Accessed 13 Jan. 2019].
