Artificial intelligence: mind games
Artificial intelligence (AI) isn’t new but the rise of mobile and cloud computing, combined with big data and cheap computing power, is driving a resurgence. Convergent technologies mean AI is finding new uses in financial services.
AI will be used in “every single segment of financial services”, predicts Christophe Chazot, group head of innovation, HSBC. “The software is getting more intelligent in a human sense, mimicking human reasoning.”
The technology can help wealth advisors, back office staff and operations, traders and corporate finance teams. Chazot says AI technology is advancing in all parts of the bank across all divisions covering capital markets, retail, corporate and private banking. “We believe AI can help us meet our three digital objectives: to deliver a better, faster service to customers, connect more clients to more opportunities and do it in a secure way.”
The bank is creating a series of labs around the world dedicated to AI, data, user experience and security, which overlap technologically and are designed to provide an in-house capability in line with some of its global competitors. The bank also hopes to mitigate the potential threat from technology-based new entrants that want to cherry-pick profitable lines of business.
Chazot says: “We need to ‘own’ these technologies by undertaking applied innovation research, co-operating with academia, and yes, by partnering with tech firms and start-ups where necessary, so that we can apply AI specifically to our problems. If we are just users, we will lose our edge.”
The threat of displacement or disintermediation is one that Josh Sutton, global head of data and AI at Publicis.Sapient, recognises, commenting: “I’d be surprised if tech companies didn’t grab some market share.”
The technologies and concepts of which Chazot speaks aren’t new – natural language processing, machine learning and neural networks have been around for years. But they are more powerful as computing power has advanced and, crucially, as the pool of available data and analytical tools grows.
Many financial institutions and technology companies such as Microsoft and Facebook have deployed powerful solutions for image or voice recognition to verify customers, use big data analysis to spot trends or opportunities and use digital assistants that are designed to help customers and automate processes. Commonwealth Bank of Australia, for example, has purchased Chip, a humanoid robot from PAL Robotics to carry out research into AI.
AI growth stagnated in the 1990s due to difficulties in maintaining the knowledge base to upgrade and manage AI systems, which was caused by a lack of data and tools at that time, says Parth Desai, the founder of Ace Software (now named Pelican). “AI has made a resurgence, however, due to the availability of vast amounts of data generated by devices and websites.”
Vijay Mayadas, vice-president of strategy at Broadridge, says AI’s time has finally come. “The rapid progress of self-driving vehicles and the defeat of the Go world champion Lee Sedol earlier this year by Google DeepMind’s AlphaGo computer program – an event that many thought was at least a decade away – are strong signals that AI is likely to rapidly gain broader adoption.”
Mayadas says significant progress has been made in applying machine learning techniques to robo-advisory and anti-money laundering and know your customer activities in financial services. “On the buy-side, many quantitative hedge funds have been at the cutting edge of implementing machine learning techniques to explore large market datasets to derive alpha. In operations technology there is also significant process automation. AI techniques can be applied to the 5 per cent of exceptions that drive 95 per cent of the work and cost of procedures.”
AI is also used to combat fraud or market abuse and battle AI-enabled cyber attacks. “AI can help detect in real time when suspicious transactions or trades are made much more effectively and on a larger scale than humans,” says Lee Beardmore, chief technology officer at Capgemini.
Nicolas Mackel, chief executive of the Luxembourg for Finance trade body is moderating an Innotribe session today (28 Sept), AI for financial services. He thinks it is early days for the resurgent AI revolution, “but Moore’s Law, which doubles computing power every 18 months, means its impact will soon be evident”. In his opinion AI is “a huge business opportunity with risk management, investment and many other end uses that will emerge in years to come”.
He adds: “I foresee co-operation between financial institutions and tech firms. I don’t think AI will displace banks, nor do I think the machines will take over.” Rather, financial services jobs will be restructured with new jobs created to service the technology and put a greater focus on customers. Financial services itself will also be transformed, he adds. “AI will blur tech and business lines and it might enable new entrants to enter financial services. I also expect to see significant one-off instances where AI causes problems.” High frequency trading and algos, for instance, led to circuit breakers on financial markets. AI may have unintended consequences.
“There is danger in every new technology,” admits HSBC’s Chazot. “We need to mitigate the risks, while recognising the opportunities, particularly how AI aligns with mobile, cloud computing and big data technology. The convergence of these trends and of applications that integrate all of them offer great opportunities that we as an industry must take.”
AI to watch
Technology companies and AI platforms to look out for are numerous but Wit.doc is interesting and was recently purchased by Facebook. The Viv AI platform enables developers to create an intelligent, conversational interface to anything. Capital One has partnered with Amazon Echo and its Alexa voice command and recognition technology to enable US customers to check their credit card balances and pay bills verbally. The Kik free messaging app for smartphones is also attracting and retaining large numbers of US teens with its add-on tools and addictive algos. If they monopolise the time of future generations what is to prevent FS applications in the future? This is the threat of closed ecosystems that leave retail banks as “plumbers” without customer facetime.
There is also the problem that investment banks might incur unacceptably high costs or lose out to more data-centric tech firms or newcomers unless they invest in their own AI capabilities. This has prompted Goldman Sachs to invest in the Kensho start-up, which is seeking to more easily and cheaply deliver the type of financial data and analysis in which Bloomberg and Thomson Reuters specialise. The bank has also invested in Digital Reasoning, which specialises in cognitive computing for enterprises and automating the understanding of human communication via its Synthesys machine learning platform.
“There are hundreds of start-ups pioneering new techniques in specific industry niches,” says Capgemini’s Beardmore. “The evolving number of use-cases and numerous projects such as Elon Musk’s OpenAI initiative mean the next five years, let alone 20, will be a very exciting time.”
By Neil Ainger, Daily News at Sibos reporter