NextGen Banking London 2019: Live blog

NextGen Banking London 2019: Live blog

Welcome to Finextra’s live coverage of NextGen Banking London 2019, the one-day event which will explore how the AI revolution will permeate and disrupt the financial services industry and what role the technology will play in the future of finance.

15:50: Before delegates head for networking drinks, Anna Milne and Gary Wright close the event and invite attendees to join Finextra again at their next event – EBAday 2019 in Stockholm!

15:46: AI is being deployed today, and is becoming more understood as a technology that is part of something else, but there is still a lot more to be done in terms of strategy, Wright concludes.

15:41: Wandhofer mentions open banking and how it has not been a focus on the corporate side.

15:26: Fahy: “Decide what you are and be good at it, - answer to question on Cynergy Bank’s AI strategy. He goes on to say that banks want five things: happy customers, happy staff, revenues, efficient costs and loans paid back. However, a bank could create a new revenue stream with AI.

15:19: How do we move from POC to pilot to production? How are incumbent banks approaching it? Dubey says that while business case has improved, it’s slow because the ROI does not have to be as strong as other project. The evolving signs means that it is still not transparent. The bigger use cases are still being evaluated.

15:14: Do regulators know enough? Or are banks behind the lack of compliance and rules? Fahy provides his experience and says that regulators don’t tell you to do something, or to not do something. “They tell you about the business risks and slow you down, until you fully understand. In relation to AI, regulators are learning their way forward because UK regulators are advanced in their thinking and embrace change. Where they don’t understand, they take small steps.”

15:10: Gary Wright encourages an interactive discussion and to provide a wider perspective. Dubey poses as a villain to the concept, one who does not believe in the technology, but has started to understand it better over the past 12 months. Wandhofer is interested in the practical applications as we move closer to the revolution, however she believes we are far from it.

15:08: Now, what of the future? Ruth Wandhofer, head of regulatory & industry at Coinfirm, Sameer Dubey, head of payments products at Barclays and Nick Fahy, CEO of Cynergy Bank join Rajiv Desai, SVP – US operations, Pelican to talk through the future of AI and B2B services.

15:00: The concluding comments for this panel session picks up on comments made throughout the event on how it is not the training data that is biased, but the people who gather this data that are biased, and this translates into the results. Williams answers a question from the audience that picks up on the point that there is not a need for diverse data scientists, but instead a range of skill sets.

14:49: Dewar is reminded of a story he once heard about Spotify and how music recommendations are made – the label was recommended before the album, but this is how our brains work, the Vocalink representative explains.

14:47: Maude: “I don’t think we will reach a point where humans will not be able to explain what it going on. We may get to a point where the cost of explainability outweighs the benefit.”

14:45: Regulators are not from a technological background – this is difficult, Williams highlights. Other challenge is looking at the outcomes and checking they are in line with what we would expect. Humans bring their own biases, but we cannot automatically test those. “Regulators have a steep learning curve to ascend.”

14:43: Maude agrees, but says that software engineering techniques like version control should be introduced – do testing where data sets are randomised before being put into the system to see how the output has changed. Provide an audit trail that regulators could start demanding.

14:39: From a regulator’s point of view, it might be beneficial for them to “unscrew” a business. Williams adds that the problem with AI is that it is not that easy, and in turn, not as easy to explain. Dewar: “Machine learning might not be at the heart of our processes, but that does not mean we shouldn’t interrogate them. We should take a look at the APIs. It’s not expensive and it’s not difficult, always able to interrogate the current status of a system and then properly regulate.”

14:36: Dewar says that there is a performance imperative that needs to be considered which can be done by measuring everything properly, establishing the KPIs. Williams introduces regulation to the discussion and with GDPR in the mix, reiterates that customers would want to know why decisions were made – and to be explained.

14:34: To what extent do we need explainability? Maude says we 100% need it for reasons outlined by Desai – we cannot just say that it is because the computer has said. People are not going to trust that answer, when they are declined for a loan application or another product. The trust that people need to have for banks to function will not be there if we leave it up to the computer.

14:31: Milne introduces Jonathan Williams back to the stage, this time on a panel with Dr Michael Dewar, VP data science from Vocalink and Jason Maude, head of technology advocacy, Starling Bank.

14:29: “Explainable and ethical AI are paramount. In banking it is essential that AI technology is compliant,” Desai says.

14:27: Using examples of Pelican data and indicators, Desai explained how as a tool, AI can recognise blacklisted areas that payments may be coming from, but in an educated way. AI understands context.

14:21: Desai highlights that both forms are important but encourages bankers to first truly understand what these technologies are. “From a system point of view, always keep in mind that explainability is important.”

14:16: There are two forms: cognitive automation, which includes payment enrichment, fraud, AML etc; or man-machine interfaces, which includes robo-advisors, chatbots or natural language processing, Desai explains.

14:10: Rajiv Desai, SVP – US Operations from Pelican kicked off the second case study of the day to discuss how to justify the decisions made in the financial services industry – again starting off with a reference to Alan Turing and WWII. “Human beings are different and no payment is the same. You need a different approach to see how decisions are made by different people.”

12:50: Breaking for lunch. Back at 14:10 for a case study on explainable AI from Pelican. 

12:42: Cordeiro: “Even algorithms need parents. And the parents have the responsibility to train them, but where are these people? They don’t exist.”

12:40: Cordeiro: AI will automate repeatable work, but where does that leave us? We could say that the workforce of the future will be more relationship-based. Banks need to look at how to foster new talent and how to develop existing teams.

12:37: Can AI be used to predict bias? Uzoma states that the issue with abuses is that they start to take on different forms, so predicting may be a little difficult. He returns to the point on education spoken about in earlier sessions and that there needs to be a recognition that we cannot look to the “altar of technology” to solve problems. Conway: “How can we make today better, than aim for the nirvana?”

12:32: As the discussion moves to regulating and standardising technology, a question was raised as to whether AI could be a leveler, as it is shining a light on all issues and the non-diverse nature of the industry. Conway agrees and gave low-level examples of what happened when testing “I’ve lost my wallet” against “I’ve lost my purse” – the latter was not recognised by the system and showed that the creators were “not thinking.”

12:25: Uzoma: At State Street, they look at how it impacts the bigger sphere and on a cultural level. How the outcomes of AI should be reflected on the customer. Conway explains that bias will be inherent, but diversity is key in avoiding it – it is not really a technology problem. However, AI also helps keep customers safe. 

12:23: Monaco also says that the right to be forgotten could become problematic, as it would also apply to institutions, not just individuals. Milne asks whose responsibility is it to ensure that the organisation stays on track. Cordeiro answers, everyone.

12:20: Monaco adds that the Commission published ‘AI for Europe’ on the 25th April. On page 10, they discuss looking at how to create a legal framework to cover this. She also mentions the EIDAS regulation.

12:16: Milne opened with a question about the scale of responsibility and Monaco says that she is glad this is being considered, as regulatory framework only exists in the form of data protection – specifically Article 22 in GDPR. This could become a source for a principle, or a new rule for AI and the use of algorithms in financial services.

12:10: Addressing AI’s ethical issues, Ekene Uzoma, VP digital product development from State Street, Monica Monaco, founder of TrustEU Affairs, Terry Cordeiro, head of product management – applied science and intelligent products at Lloyds Bank and Michael Conway, associate partner, global business services at IBM come together for the third panel session of the day.

11:59: Hardie covers the results – a lot of fine tuning resulted in false positives being eradicated. £12.5 million recovered from customers such as charities and schools, reinforcing the relationship with the customer after having discussions about potential fraud.

11:53: 2.2 billion payments are processed by NatWest every year, and less than a few hundred cases of fraud. A small needle in a big haystack. After partnering with the NatWest Fraud team, Vocalink worked with the transaction data to develop the machine learning model, fine tune it to then score the frauds.

11:51: Marc Corbalan returns to the stage, but this time with David Hardie, fraud prevention manager from Royal Bank of Scotland, for the first case study of the day discussing tackling financial crime and the role of digital identity.

11:44: Reid: AI solutions depend on significant volumes of training data – so they may need to be retrained and the lifecycle becomes different. Jain on scalability – “Today, we haven’t mentioned the devops capability and those who are aware of business problems, and a second point needs to be made on bias and the importance of diversity of views when creating the model is needed.”

11:37: Janusz on regulation – “it comes when something crashes. Banks will be reluctant to implement AI without human oversight.” Reid adds: AI is good at tackling problems where you can’t apply step-by-step rules. AI works best in an uncertain space and it’s beneficial to reflect back on what the technology is good at.

11:29: Reid: “The hype around AI has created the impression that there is such a thing as an AI solution, but AI plays a part in the products and services that we offer, and in the future, will play a part in everything we do.” Akerkar also points out that the most value that can be gained from AI is when you think of it as stages of decision-making.

11:26: Janusz questions how banks can evaluate the ROI of this “discovery process” if they do not know what the end result will be. However, the learnings may be more valuable than the final delivery.

11:22: Jain introduces the subject of maturity and discusses how costs, as well as the technical aspects of a project, must be considered. Akerkar says that the data science lifecycle is different to the traditional IT product lifecycle, but with new technology, it is a “discovery process, but you are still trying to figure out where you will end. You will learn something after the POC or pilot of an AI product, rather than just ticking a box.”

11:15: Akerkar answers Williams’ question about how to promote projects, and flips it on its head and states that it is important to understand the technology and the business needs before the use cases. Reid follows up and says that from the vendor side, if the technology is poorly understood, then it might be a good idea to “try it out” and then see what it can do practically.

11:13: Jonathan Williams welcomes Maciej Janusz, head of cash management Nordic region, Citibank, Dan Reid, CTO and founder, Xceptor, Abhijit Akerkar, head of applied sciences, business integration, Lloyds Bank to the stage, with Jain following his keynote.

11:08: Fintech partnerships and AI – “there’s a lot of AI/data/ML-type consultancies so when you want to expand, you’re not competing with Amazon and you can get your human capital from there.” How do you create a partnership where all parties can learn? How many of you interact with developers? – Jain asks. All of this only works if your organisation is on board. “AI is hyped up and is at the peak of its curve, but it is real and it is practical.”

11:03: “Banks were established to be trusted parties, but now we’re seeing the monetisation of data which is having an impact,” Jain says. Where are we seeing this impact coming from? 1. Within the organisation, 2. Within the industry and 3. Talent – the latter which we should think of as “human capital”. Existing employees looking to do things differently are your best asset.

10:57: Jain focuses on how a lot of discussion about AI is around FOMO – fear of missing out – but there is nothing new about it, machine learning was used in WWII. What has changed is customer expectations and consumers wanting their banking services to be available in a couple of clicks. Margin erosion and the workforce of today are also big triggers.

10:55: Anna Milne summarises the morning’s notes before Karan Jain, head of technology Europe and Americas from Westpac, takes to the stage to discuss specifically how AI is disrupting the financial services industry.

10:13: Time for coffee, join us back here at 10:55!

10:10: Bannister concludes with a comment on how AI has gone from hype to reality, and now we’re comfortable with the technology and we know it will go ahead. What’s next?

10:04: Rohatgi: “The real knowledge rests with the few – there are a lot of tools, but the business doesn’t know how to use them, but that’s not their job, their job is to run the business. The knowledge needs to be shared. The dichotomy that exists is between the expert and the business.” Orritt adds, “the computer says no.”

10:00: Adams: “10% of the AI project is AI – the easy bit. 40% is the data, where you’re getting it from and cleaning it and 50% is the regulatory risk.”

9:58: “There are so many industries that will be impacted by AI, but we always overestimate the technology change in the first year and underestimate it for the decade,” Adams said, referencing the Apple iPhone. The right risk frameworks need to be put in place, because robotic products can be quite sensitive.

9:55: Will jobs become more satisfying? Adams sees the lives of everyone being enhanced because of these technologies, “as long as our leaders focus on diversity and inclusion.” Orritt adds that Santander looks after their vulnerable customers – “we’re not going to send a robot to have a difficult conversation, there always needs to be a human layer.”

9:50: Adams: “There is a danger of the gender pay gap getting bigger because of AI.” She discusses how attitudes are different in China, India and Romania, so girls need to be educated and given the opportunity to apply for jobs in data.

9:48: Rohatgi agrees with the education point, as there is a lot of confusion around what AI actually is. “Everyone is keen to use this stuff, but the system, the fabric, is not mature yet. It’s all well and good to go from POC to pilot, but it never really reaches the real world.”

9:46: Adams talks a little about the benefits of educating employees within financial services, but hones in on the point that other than truly understanding the technology, investment in AI must be done efficiently, and only then, strategic advantage can be gained. “Data equals training equals insight.”

9:41: Orritt: "There’s a lot of hype within the banks and in the industry and it’s about taking it off the pedestal. Deep learning is only going to solve 5% of your problem," he said, as a response to Adams' comments in her keynote.

9:39: David Bannister welcomes Jonathan Orritt, data scientist, Santander, Roshan Rohatgi, artificial intelligence lead, RBS and Janet Adams back to the stage for the first panel session of the day. Rohatgi starts off by explaining that there is a lot of exploratory work to be done, but there is an opportunity.

 

9:36: Corbalan: “AI is a tool for now and we expect much greater reliance on this technology in the very near future. The AI world has benefited from the collaborative framework.” He goes on to mention how this AI revolution is only possible today because of computing power" – some food for thought.

 

9:33: Another example Corbalan provides is how machine learning can be used in forecasting and predicting future revenue profiles. ML here allows for a “step change in accuracy and reducing the need for manual forecasting.” A second way AI is used here is for labelling.

9:29: Focusing on fraud, Corbalan points out that Vocalink uses automation to figure out whether fraudsters are human, or are themselves automated processes (robots) – and adds a note on how impactful this could be for open banking, especially if multi-layered scams occur.

9:25: In Keynote 2, Marc Corbalan, product management at Vocalink, highlights that AI will transform banking and is already moving the needle today by solving business challenges that previously could not be solved. He goes on to discuss how Vocalink are able to leverage payments data points to build analytic services.

9:22: Adams feels as if the industry is at the point where there is consideration about the integrity and ethics of AI, and links this to the potential impact on the Principles of Business. “The AI revolution isn’t happening in a vacuum,” but in banking, it presents an opportunity for us to look at how we can be a better industry.

9:17: Why are neural networks so important? Adams explains that it can be seen as an IQ stack and when you’re speaking with your peers, they will be discussing things like traditional AI, which includes Fuzzy Logic, Random Forest etc, but only when you get into neural networks, or deep neural networks, AI beats humans. "Deep learning has got everyone excited and has put AI on the map. No one has found deep learning’s limit to learn."

9:15: Adams: "AI is the new electricity – it powers everything and will be a huge part of every job that we do. We’re standing at the brink of this wealth creation. TSB Bank aim to help their customers through the wealth creation stage of their lives." 

9:10: Janet Adams, head of AI at TSB Bank, started off her keynote stating that she read that AI would disrupt across all industries and presented a video which saw a robot collecting millions of data points. She explained that the way in which machine learning and neural networks work is with "weights" and the results are compared to the robot’s – this is done hundreds or thousands of times, depending on the computing power. “That is the foundation of machine learning and machine learning in financial services,” Adams said.

9:06: Finextra editor Anna Milne and content director Gary Wright kicked off proceedings, welcoming a full room of delegates to Finextra’s second event dedicated to artificial intelligence and encouraged interaction. Wright added that the two words that stand out are ‘digital’ and ‘data, more on that across the rest of the day.

7:30: Good morning and welcome to Finextra’s live coverage of NextGen Banking London 2019. There is a packed programme to look forward to today, scrutinising AI’s promise to impact businesses and become a strategic priority in the financial services sector and beyond. Proceedings will kick off at 9:00am, so check back then!

Editorial | what does this mean?
This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community.

Comments: (0)