Conceived in the 1950s when the Logic
Theorist became the first ever program to mimic human problem-solving skills,
artificial intelligence (AI) is today advancing at an unprecedented speed. A
breakthrough moment arrived in March 2016: Google’s AlphaGo beat Lee Sedol 4-1
in a five-game match of Go, the first time a computer program beat a 9-dan Go
professional without handicaps. (AlphaGo had been created in 2014 to test how
well a neural network using deep learning could compete at Go.)
To understand how we got here, look no
further than the AI triple play: data, computing power and cognitive
science.
Taking these components in turn: firstly, data, the raw material of
the information age, has grown exponentially; between 2002 and 2015 global
internet traffic volume grew from 100 GBps to 20,235 GBps [1].
Secondly, the computing power available at our fingertips is now astonishing – a smartphone today is more powerful than all the computing power NASA had in
1969 when Neil Armstrong became the first person to walk on the Moon.
Thirdly,
advances in cognitive science mean our understanding of how the brain works is
better than ever before – which in turn helps to frame deep-learning programs
in their efforts to simulate human decision-making.
Taken together, this is a huge opportunity.
Our primary challenge is to better grasp how we can harness the vast power of
AI – and adapt to the new world it will create.
Historical internet context
| Year | Global internet traffic |
| 1992 |
100 GB per day |
| 1997 |
100 GB per hour |
| 2002 |
100 GBps |
| 2007 |
2,000 GBps |
| 2015 |
20,235 GBps |
| 2020 |
61,386 GBps |
Source: Cisco VNI, 2016
The opportunity for banks – and for clients
Banks are at heart large, regulated technology companies: they produce, manage and receive vast amounts of data – only a small portion of which is well structured and channelled through data-modelling techniques. The remaining bulk of unstructured, underused data is where the opportunities are: it is where AI must take aim. (A universal bank like BNP Paribas, which has a full range of clients from individuals at retail level to large-cap corporates and financial institutions, has even more data to aim at.)But how? Many algorithms are now available open-source, offering opportunities for developing new solutions internally – such as translation, web crawling, name and image recognition, chatbots and natural language generation, to name but a few. This means the algorithm is no longer a differentiator. Instead it is the relevance of the algorithm that is the differentiator, i.e. what determines the relevance of the algorithm is the dataset at which it is aimed.
To this end, translating data into opportunities and better services for clients can be achieved in four ways: first, we have to work smarter, work better, and work in partnership with our clients; second, we must adapt existing IT infrastructure to be more modular and more open; third, we must reinvent how businesses and IT departments work in partnership with each other; and fourth, we should explore all AI competencies, whether in partnership with start-ups – or internally.
Man (and woman) and machine
Just as chess-playing computers were fashioned by a combination of people and AI, the most successful deployment of AI will be when humans and learning machines form a symbiotic relationship: we should see AI as complementary to human activity. With the right framework in place – widely available, interoperable APIs set on the right bits of data, partnerships with clients, etc. – AI can help people become:
- more efficient in executing long, difficult and costly tasks;
- more powerful in performing large-scale analysis;
- smarter in selecting options;
- faster at making decisions on data sets; and
- better at anticipating and detecting opportunities
In sum, AI can help banks like BNP Paribas be more relevant and more useful to their clients. AI can help us to protect our clients’ interests, anticipate what they need and suggest new investment opportunities; and it will mean we will have to reinvent ourselves, redeploying the time freed up by AI to focusing on customer experience and customer satisfaction. We are, after all, emotional beings.
People will always be needed as a trusted partner for clients – above all to foster responsible, sustainable growth.
[1] Source: Cisco, VNI