Pauline Norstrom is a C-level executive board director. She has 25+ years experience in leadership including MD/COO/VP/President/Chair roles, developing and transforming international “tech” teams and companies in video surveillance, cyber, AI, IoT domains at all stages of the life cycle. A key interest is in identifying market opportunities for AI and effectively executing strategy across multiple verticals including transport, retail, banking, education, utilities, process engineering.

I had a very interesting conversation when we met a few weeks ago around her background and emerging technologies, including AI (artificial intelligence), that are becoming more prevalent in many industries and our day to day lives.

Justin - When did you first become aware of AI and how did it apply to your work at the time?

Pauline - I have been aware of various forms of AI for nearly 20 years. In my work, I have tracked the development of AI in image analysis in the video surveillance domain since the early attempts in the 2000's to improve detection, search accuracy and speed.

Prior to the emergence of image analysis, specialized video surveillance devices produced outputs resulting from a varying range of conditional events. Although not described as AI at the time, the automation of a task using variables which are conditional on the environment is a form of machine learning which is now categorized under the AI umbrella.

In my roles on the Boards of leading technology companies over the last decade, I wished for AI to be available to crunch together structured and unstructured data from multiple sources to enable better decision making.

Justin - What most excites you about AI?

Pauline - When used ethically, transparently and within appropriate legal frameworks, there is great potential for AI to add value by freeing people from simple repetitive activities delivering knowledge rather than just data to enable better decisions.

The use of AI to improve diagnosis and outcomes in healthcare and the use of facial recognition to automatically and quickly find people in crowds could improve the health, security and well-being of millions of people.

The advances in the core technologies such as 5G and Quantum computing open up the accessibility and speed of AI to make it available as a converged service for the benefit of society and industry by bringing together otherwise disparate systems.

Justin - AI is a bit of a buzz word right now, how do you see AI helping in your industry?

Pauline - The term artificial intelligence holds a different meaning to different people. Whatever the perspective on AI, it is a term used to re-classify a number of technologies including machine learning and deep learning.

Deep learning technology improves detection and search functions in video surveillance so a person can be differentiated from some other object. Relevant video and data is automatically and securely combined using AI to deliver relevant information rapidly to the right people.

The suite of AI technology available to the industry improves situational awareness leading to the fast deployment of emergency services to take more directed and appropriate action. It is also enabling improved productivity and management of smart building and cities.

Justin - What do you see as the main challenges to rolling out an AI based solution? We spoke a lot about internal challenges with company stakeholders, and external challenges with the general public’s perception of AI and how data is used.

Pauline - Boards need to acquire skills and knowledge to enable them to assess the relevance and need for AI in their organisations. AI can affect the entire business and its culture. It can make or break reputation, the perception of the public and can affect shareholder value. It can also bring substantial benefits to a business if the strategy is led from the top and ownership and accountability is with the Board.

This means that the makeup of the Board should be examined and consideration given to “skilling up” or taking on an NED or NEA to advise on the use of AI, and it certainly should grow a place on the Board agenda.

The term AI can be synonymous with change and for many, the idea of change can involve fear of the unknown. It is very important that there is transparency within the organisation to stimulate feedback and debate and as a result there is a higher chance of the project being successful.

External stakeholder management involves really good communication which explains the technology, its benefits and the potential uses. Success stories provide proof of safety and effectiveness helping to bring the public along on the journey.

Justin - There are some very exciting start-ups developing innovative AI based solutions. Sadly, only a relatively small percentage will get their product to market and be commercially viable. When working with companies that are launching AI initiatives or start-ups that are developing AI products, what are the most common areas they need support in or need to fix?

Pauline - The AI start up is often driven by the technical teams working in an intense development environment. A balance has to be struck between the commercial and technical teams, respecting the creativity of the developers whilst ensuring that a minimum viable product (MVP) which solves the customer’s problem has been specified and a decision made to launch with an agreed feature set. A delay to market while this is debated, can leave a gap for a competitor to fill whom may copy the new technology quickly and enter a pre-prepared space which is ready to adopt it.

The development of the value chain which recognises the players whom can form part of the route to market it an essential element of a successful launch.

Whether AI start-ups will succeed or fail is dependent on having the right people in place, skilled and experienced in leadership, business and technology. Thorough research, understanding of the customer need and discipline in defining the MVP are critical elements.  When innovating in AI, there are millions of potential variables and it is not always known whether the product will operate in the way intended, therefore sufficient time should be allocated to attracting early adopters and testing in the field prior to a roll out.

Justin - The EU is looking to introduce rules around the use of facial recognition technology. How can we reach a balance between data privacy and the use of facial recognition in preventing terrible events like terror attacks?

Pauline - Both EU and UK are looking at legislating beyond the GDPR. Currently facial recognition is being used in public places in the UK by The Metropolitan Police following the guidelines laid out in this current legal and ethical framework.

What may not be known by the public is that facial recognition technology is already extensively used in education, border control and retail.

To reach a balance between data privacy and protecting the public, the video surveillance and security sectors explain how the technology works and why it is used and for what purpose.

The majority of the 6 million plus video surveillance cameras in the UK are owned and operated by private entities not the Government. The ethical considerations towards the use of potentially intrusive technology is woven into the fabric of all relevant standards which security companies are required to work to if they wish to maintain vital certifications and their status as quality providers in the marketplace.

The industry welcomes specific legislation and regulation underpinned by standards which keep up with the pace of development while still protecting the privacy and rights of the individual. Where the application of facial recognition can be regulated within current legislation there is a case for its continued use as its removal may be a backward step which results in an increase in crime and loss.

Justin - If you had one wish and could change anything about AI, its perception, data laws, government support etc what would it be?

Pauline - To reassure the public that AI should not be feared instead to consider it to be vital element of society.  Its cross examination with regard to bias may actually lead to the reduction of bias in society.

I would like to see the continuing drive towards open source technology and the creation of vast, diverse and balanced data sets representative of society and how we would like society to evolve. The emergence of AIs which test other AIs for bias may help to demystify the algorithmic decisions.

It is clear to me that the matter of ethics in AI means different things to different industries, for example the medical profession works to a different ethical code to that of financial services. One size does not necessarily fit all therefore the development ethical codes and the explanation and use of the ethics should developed through the collaboration of Government and the specific industries.