Problem or Potential? Considering Ethical Artificial Intelligence in the Financial Sector
Artificial Intelligence is being used across multiple industries worldwide and is seeing dramatic growth. Annual investment by Venture Capitalists into US AI start-ups has seen a six-fold increase over the past 17 years.[i]
The financial sector is one industry which can enjoy multiple benefits from the utilisation of AI. Statistics suggest that over the next three years, AI could decrease bank’s costs by 3.9%, while increasing revenues by 3.4%. The dawn of AI in the sector is good news for customers too, as AI can improve processes for savings, investments, risk assessments and loans; in fact, across almost all aspects of financial services.
However, with the rise of AI, has come concerns about the ethics surrounding its use. Worldwide, there has been much debate about how to develop and use AI in a responsible way. The House of Lords has produced a report on the ethical use of AI, highlighting a number of principles that must be applied to the development of AI.[ii] This includes the following:
The common good
The House of Lords Select Committee highlighted the importance of all AI development being for the common good of humanity. Though AI focused on improving healthcare, education and the environment may seem to be more in line with this mandate for “good”, there is no reason that the financial sector cannot also have a thoroughly ethical framework for AI development. AI within financial services can be focused on creating systems, processes and advice that are accurate and helpful. AI can give immediate responses to queries and offer instant transactions. It can help to identify anomalies, thus potentially exposing fraud or error. Indeed, a government review identified the three areas with the most potential for AI to be in anti-money laundering and fraud detection, as well as personal financial planning.[iii] The aim should be to create an AI that helps its customers to take the course of action that is most beneficial; with none of the human error or bias that may come with a human advisor.
Fairness
It was established that AI should be developed on systems of fairness, being comprehensible, user friendly and accessible. The financial sector must focus on developing AI that will be intelligible, transparent and easily understood by clients.
A new chat bot, currently being developed by AiX, can analyse the input of traders and participants, working essentially as a financial broker, but with the potential to reduce brokerage fees by an estimated 50%. Conversations will be recorded so that it is possible to audit the financial advice that it gives; it will be possible to clearly evidence the train of logic that was used for the chatbot to quote a particular price. While human interaction needs to take account of personality, tone of voice, humour, or sarcasm, all of these idiosyncrasies are removed with the chat bot, leaving transparent, bias-free financial advice.[iv]
Privacy
With respect to data and privacy, new AI systems must not diminish people’s rights. By its nature, AI needs to gather lots of information about an individual; their habits, preferences and finances in order to offer the best service. This is of benefit to the client, but also poses a risk in terms of protecting their data. However, these risks are not insurmountable. Accountability is vital as systems are put in place to ensure privacy for individuals. Keeping in line with recent GDPR legislation should give some good guidelines in terms of protecting people’s data while developing AI. AI may even be able to further privacy requirements; for example, by anonymising data at the source. Much research and consideration needs to go into the area of privacy and data protection when developing AI in the industry.
Education
The report also specified that part of establishing ethical use of AI, involves educating people to be able to flourish alongside AI. Already we can see that the percentage of jobs that require a skill of working with AI is 4.5 times greater than those in 2013.[v]
Within the financial sector, there is huge scope for human advisors to work with and alongside AI to provide the best service to customers. An AI can make a human advisor eight or ten times faster in their responses.[vi] While AI may be able to overtake even the most expert human counterparts in data analysis, accounting and planning, people still want to speak with another human when discussing something as important as their financial future. AI can aid the human workforce, rather than taking over from it, but companies need to be pro-active about training their teams and offering outside training, as well as petitioning educational institutions to cover modules on AI in finance.
Specialist recruitment agencies such as Aston Charles are well placed to be able to provide financial companies with job candidates that have the skills and training to work alongside AI, driving the company forward as they face (and embrace) this disruptive new technology.
Is AI unethical?
Any sector would be naïve to suggest that there are no ethical concerns within the development of AI. However, with the right ethical framework, it should not be an insurmountable problem to develop AI that has a highly positive impact on the financial sector, making it fairer, more transparent and offering huge potential for companies and their clients.
[i] LeGassick, C. (Project Manager), 2017. Artificial Intelligence Index. 2017 Annual Report. 3rd ed. Stanford University: Available as PDF. Page 17.
[ii] UK Parliament. 2018. UK can lead the way on ethical AI, says Lords Committee - News from Parliament - UK Parliament. [ONLINE] Available at: https://www.parliament.uk/business/committees/committees-a-z/lords-select/ai-committee/news-parliament-2017/ai-report-published/. [Accessed 18 May 2018].
[iii] Hope William-Smith. 2018. Will AI Change the Face of Financial Advice. [ONLINE] Available at: https://www.moneymarketing.co.uk/will-ai-change-face-financial-advice/. [Accessed 20 May 2018].
[iv] Business Insider. 2017. Banks are excited about potential of AI in finance - Business Insider. [ONLINE] Available at: http://uk.businessinsider.com/ai-in-financial-services-2017-11. [Accessed 19 May 2018].
[v] LeGassick, C. (Project Manager), 2017. Artificial Intelligence Index. 2017 Annual Report. 3rd ed. Stanford University: Available as PDF. Page 18.
[vi] Hope William-Smith. 2018. Will AI Change the Face of Financial Advice. [ONLINE] Available at: https://www.moneymarketing.co.uk/will-ai-change-face-financial-advice/. [Accessed 20 May 2018].