The Difference AI Can Make

McKinsey have issued a white paper on what is happening with AI and Generative Pre-trained Transformer (GPT) ¹. Key take-aways were:

Impact on productivity: McKinsey analysed 63 use cases drawn from 16 business functions and found that its impact on productivity would be measured in the trillions of dollars. About 75% of that value would come from customer operations, marketing and sales, software engineering and R&D.

Banking, high tech, and life sciences could see the biggest impact as a percentage of their revenues from generative AI. In retail and consumer packaged goods, the potential impact is also significant at $400 billion to $660 billion a year.

Generative AI could enable labour productivity growth of 0.1% to 0.6% annually through 2040, depending on the rate of technology adoption and redeployment of worker time into other activities.

Combining generative AI with all other technologies, work automation could add 0.2 to 3.3 percentage points annually to productivity growth. However, workers will need support in learning new skills, and some will change occupations.

Automation of work: Current generative AI and other technologies have the potential to automate work activities that absorb 60% to 70% of employees’ time. Previously McKinsey had estimated a 50% change, but the ability of generative AIs to understand natural language, which is required for work activities that account for 25% of total work time, has led to this increased estimate.

In addition, generative AI will have more impact on knowledge work associated with occupations that have higher wages and educational requirements than on other types of work.

McKinsey estimates that half of today’s work activities could be automated between 2030 and 2060, with a midpoint in 2045. Technology development, economic feasibility, and diffusion timelines are accelerating.

Mastercard Uses AI to Tackle Scams: Mastercard has been working with UK banks to follow the flow of money mule accounts for a number of years. It is now combining this with other criteria, account names, payment values, payer and payee history, and the payee’s links to accounts associated with scams, to give banks intelligence to intervene in real time and stop a payment before funds are lost.

Nine lenders are using the technology in their fight against APP fraud which accounts for 40% of UK bank fraud losses.

Risks of AI

A different side of AI was shown in the chief of the UK’s Financial Conduct Authority warning banks about the risks from the rise of AI and the threat of ‘deep fake’ fraud. His remarks were mainly concerned with the use of AI bots in automated trading and of unwanted outcomes from biased datasets.

The FCA is concerned about the risks that Big Tech may pose to operational resilience in payments, retail services and financial infrastructure. An area of focus is the risk that Big Tech could manipulate consumer behavioural biases. Big Tech firms have a role as gatekeepers of data, and this has implications of the ensuing data-sharing asymmetry between Big Tech firms and financial services firms.

There are competitive implications if Big Tech firms have access to unique and comprehensive data sets such as browsing data, biometrics and social media. If these are coupled with anonymised financial transaction data, over time this could result in a longitudinal data set that could not be rivalled by that held by a financial services firm. Because of the global position of Big Tech firms, the data is likely to cover many countries and demographics.

The ‘deepfake’ video risk comes from the ability of AI to mimic language, audio and video. A deepfake video was recently used in the UK featuring a well-known and trusted personal finance campaigner supposedly selling speculative investments.

Regulatory challenges

In 2022 75% of UK financial services companies used at least one of machine learning, deep learning, and high-performance computing, according to NVIDIA., these being AI use cases. 30% of UK bank job adverts include AI.

Financial services organisations are continuing to use traditional forms of AI but are also exploring and investing in generative AI. Bloomberg has introduced BloombergGPT, a 50-billion parameter large language model, purpose-built from scratch for finance.

All this brings regulatory challenges. Traditional AI can be used for first order analysis, rules-based sorting, classification, and Q&A engagement, but it’s not sufficient to cover all edge cases, respond to never-before-seen problems or exercise judgment. However, generative AI is.

Generative AI is a completely new platform offering a new way of communicating with technology. Its ability to simplify complexity means it has the potential to allow financial service bots to serve customers better, to make them ‘emotionally available’ to customers.

Today’s governance and compliance requirements mean the customer experience is often high cost and low satisfaction. Generative AI should reverse this if it is done well. It should allow many of those who are currently financially excluded to be included.

The UK’s Financial Conduct Authority (FCA) has introduced new consumer protection rules that will allow them to punish harm done to customers from AI systems. The FCA’s ‘Consumer Duty’ sets higher and clearer standards of consumer protection across financial services and require firms to ‘put their customers’ needs first’. In the Autumn 2023 the UK government will host the first major global summit on AI safety.

In Europe, the AI Act is the first piece in a series of complex legislation designed to improve transparency around its development, but which may slow its use. In the US, by contrast, there is a much lighter touch approach.

1 - www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-AI-the-next-productivity-frontier