DIGITALEUROPE: AI as an opportunity for Europe

A concept image of AI
© iStock/chombosan

Cecilia Bonefeld-Dahl, Director General of DIGITALEUROPE, met with SciTech Europa at ICT 2018 to discuss AI in the European context.

IN December 2018, SciTech Europa travelled to Vienna, Austria, to attend the ICT 2018 event. This research and innovation conference attracted 4,800 visitors and focused on the European Union’s priorities in the digital transformation of society and industry. It presented an opportunity for the people involved in this transformation to share their experience and vision of Europe in the digital age.

Speaking at the event, Cecilia Bonefeld-Dahl – the Director General of DIGITALEUROPE (an organisation which sits on the editorial board of SciTech’s sister publication, Government Europa Quarterly) – delivered an address in a session on artificial intelligence (AI). This session was designed to inform delegates about the EU approach to AI and to engage in a debate on the main opportunities and challenges brought about by these technologies.

SciTech Europa met with Bonefeld-Dahl on the sidelines of the event to discuss some of these areas in more detail.

What would you say are the biggest barriers and opportunities presented by AI in Europe?

The opportunities presented by AI are almost endless. However, there are perhaps two areas in which Europe should focus on in order to take advantage of them. The first is the fact that we have strong public sectors, but we have not even started looking at what AI can actually do in this sector at scale.

In the health sector, AI could steer Europe towards preventative care; or it could be employed in the development of greener environmental practices, to predict energy usage, enable smart traffic flows, optimise engines, and so on.

Second, there is a real need to look at the benefits for the education sector. We need to develop a way to forecast what the job market will be like in the future, infer what skills will be required, and then ensure that these skills are being learnt so that the next generations are not being left behind amid digital transformation. This is starting to happen, and, indeed, we can already tell our children which skills they may want to develop if they want to enjoy high-earning prospects.

This will also be important for Europe’s businesses. Companies need to be aware that in the near future they may need to relieve a portion of their workforce due to skills redundancy. As such, these businesses would clearly benefit from being able to accurately forecast labour market needs in terms of skills.

What are the biggest barriers to a wider development and uptake of AI by the electronic components and systems (ECS) industry in Europe?

Deployment is a clear issue, as is the cultural barrier that is feeding concerns about what could happen if AI ‘went wrong’. However, people should embrace the future with confidence and optimism, and we must focus on the many solutions and opportunities created by AI. It is our collective duty to define what we want rather than what we do not want. We need to first ask ourselves whether we want our children to have the right skills, which will enable them to succeed in the job market of the future; whether we want them to live in a cleaner environment where we save energy; and whether we want them to be able to access preventative healthcare and live longer, healthier lives. If we do, then we need to start realising the full potential of AI.

With this in mind, it is important that we continue to consider citizens as the most important asset at hand. As such, human-centric AI will play a fundamental role moving forward. For this to happen, however, more questions need to be answered:

• What is human-centric AI?
• Other than being unbiased, what else does it contain?
• What does it do?
• What is the concrete action?

We must always remember that AI is just a tool; we can use it to either harm or to improve today’s society.

A further opportunity resides in the fact that Europe as a society is focused on not projecting biases. While humans are the most structurally biased animals in existence, AI can offer the possibility of eliminating this in decision-making. Indeed, people attending job interviews today can experience bias because of their gender or the colour of their skin; they can be discriminated against by a single person or a group. In the future, however, this problem could be solved through the implementation of AI containing standards that have been developed to ensure non-bias. This has the potential to rise above human limitations and allow for fact-based decision-making across contexts.

Of course, this will require the development of a standard that will allow us to include a non-biased quality within the AI, and this requires a discussion around ethics.
When defining guidelines on ethics – as well as those in other areas – a principle like non-bias is applied to a standard for the development of AI. In Europe, this is something we are particularly good at. For example, while final decisions about the flying of an aircraft may be taken by the pilot, it is the AI that computes and decides the preferred altitude if a storm is encountered. This holds true for many other scenarios, but it is also something that is not further considered. Indeed, the positive benefits of AI need to be enshrined in a shared mindset to be applied to different contexts. For instance, sorting through requests for a financial loan, or ill persons needing treatment. This will undoubtedly contribute to an even more intelligent decision-making.

A significant barrier to this is the fragmentation present in Europe as there is no uniform market. Again, the culture of concern and fear around the potential harm that AI could bring continues to prevent it from achieving its full potential.However, a common strategy is now being discussed by Europe’s leaders – a significant step in the right direction. It shows a trend that Member States are not working in silos, which is crucial if Europe is to compete in the global economy.

What should be done to ensure an appropriate ethical and legal framework?

The first steps are already being taken. Europe has a tremendous amount of experience and expertise in setting standards for technologies. At the moment, work is underway to define the factors required for non-bias, which can be built in as check points in the development phase.

An important element concerns intent. AI, of course, can learn by itself. While there is original intent behind any product when it is developed, AI technologies have the potential to dramatically change throughout their life. It is important to review these technologies, looking at what the original intent of the algorithm was and whether this is still being delivered.

Standard-building is something that everybody wants. If a developer or manufacturer is aware that they need to track intent and include traceability, as they are accountable, then every industrial sector will be willing and able to do that. Industries such as pharmaceuticals, aviation, automotive, and manufacturing have already implemented this. We need to expand AI on this basis, rather than trying to reinvent the wheel, which would be a mistake.

Is there a need for the establishment of a European infrastructure for AI?

Yes, because capacity building brings obvious benefits. The more we enable businesses to scale-up, the more attractive Europe will be for companies to invest and establish themselves. Currently, Europe has just 11% of the world’s unicorns (6% of which are in the UK, meaning that post-Brexit Europe will be left with only 5%, which is pretty worrying). Consequently, capacity building will be crucial moving forward, as will ensuring that any support is easily made available to those who want it, otherwise scale-up companies will be looking to settle elsewhere.

Is there enough support at the European level for scaling-up companies – as well as perhaps helping spin-offs and start-ups?

The Digital Europe Programme is now being finalised by the European Commission and it is emerging as quite a viable platform. Indeed, while small in scale it certainly represents a step in the right direction – not least because, in the past, the EU has been perceived as providing financial support to dying industries rather than projecting visionary approaches.
Capacity building in the digital space might mean that some industries, such as agriculture, can be taken to a different development level where they can build common platforms and handle cross-border exchanges. The Digital Europe Programme is not only a step forward for the manufacturing sector but also for other, more declining sectors. Rather than providing financial assistance to such sectors to counter their losses and provide life support, it is now possible to inject funding into endeavours that can actually facilitate a shift towards a digital economy.

Commissioner Ansip has said that Europe needs to invest at least €20bn in AI by the end of 2020, while in a presentation today DG CONNECT’s Khalil Rouhana has said that the overall ambition is significantly higher. Is this achievable?

I think so, although there is a long way to go – the money that has been earmarked at the moment for AI is just €8bn, half of which will come from industry. There are two factors which are important to note here. The first is that the money will also be spent on skills. At the moment, the digital economy thrives where the best skills are. Consequently, putting money into building skills will attract foreign capital.

But here too there is more progress that needs to be made. There is a sense that Europe’s Member States need to realise just how important this is. Although we are seeing several European leaders, such as France’s Emmanuel Macron, Germany’s Angela Merkel, Denmark’s Lars Løkke Rasmussen, Finland’s Juha Sipilä, and Sweden’s Stefan Löfven, beginning to realise that the more they invest in, the higher the return.

How can DIGITALEUROPE help foster in this necessary and new environment for Europe to take full advantage of the potential offered by AI across sectors?

We are active in three areas:
• Investment
• Skills
• Regulation and guidelines, including codes of conducts and soft and hard regulations

Regarding the latter, we will continue to provide advice on how to deregulate, or how to revise legislation that should ensure Europe is able to achieve its goals with regards to AI. Additionally, we can also advise on where new legislation might be needed. Any new legislation needs to be evidence-based, perhaps more so than in the past. This will require approaching development in a more collaborative way.

In parallel, it is also important to develop much shorter processes, which will also ensure that we are able to minimise any unwanted consequences when working to achieve what Europe wants and needs.

Cecilia Bonefeld-Dahl
Director General
DIGITALEUROPE
Tweet @DIGITALEUROPE
https://www.digitaleurope.org/

Laboratory Supplies Directory - Now Live

LEAVE A REPLY

Please enter your comment!
Please enter your name here