robot Intelligence artficielle et usage responsable des données

Published on 29 December 2020

Ensuring responsible artificial intelligence and data us

Artificial intelligence is touching ever more areas of our lives, from conversational agents to work, heath and manufacturing. However, its promises also raise questions in terms of responsibility and ethics.

 

 

In line with our purpose, we believe that AI can only truly serve individuals, industry or society in general if it helps everyone. This vision translates into our commitments, research and development, training initiatives and innovation.

 

What do we mean by artificial intelligence?

Artificial intelligence (AI): définition

The term artificial intelligence is often used to describe machines (or computers) that mimic cognitive functions associated with the human mind. Capabilities can include learning, language comprehension, sensory capabilities, decision-making and problem solving.

 

 

AI and data form one of our ambitions

Increased use of AI and data across the business means we can improve our customers’ experience, make our networks smarter and become more efficient and agile.

Steve Jarrett, Vice President Data and AI at Orange, explains why and how we’re placing AI and Data at the heart of our innovations.

 

Optimal AI relies on high quality data

To work best, whether in transport, health, the environment, agriculture or industry, AI relies on processing massive amounts of data coming from a range of sources. Algorithms are used to process the data and design more efficient and useful solutions.

The massive production of raw data (Big Data) is a global phenomenon: volumes are increasing by 61% each year and will continue to do so until 2025. It’s up to us to make sense of it. 

That’s why we’re focusing on:

  • developing AI within an ethical and transparent legislative framework
  • ensuring we have the right governance and processes to supervise how data is collected and used
  • verifying data is targeted and of high quality to avoid bias
  • respecting our customer’s privacy and protecting their personal data 

 

AI at a European level

AI and data are complex issues but not globally regulated. Orange is one of the 52 independent experts on the European Commission’s High-Level Expert Group on Artificial Intelligence (HLEG AI). Together, the panel has defined seven key requirements for AI systems: human agency, technical robustness, privacy and data governance, transparency, societal and environmental wellbeing and accountability.

AI patents

 

AI must be developed as part of an ethical and responsible approach

In France, we actively contribute to ImpactAI, a think and do tank that brings together different stakeholders: companies, private or public entities, research institutes and educational partners around two common objectives for AI: address ethical and social challenges and support innovative and positive projects for tomorrow’s world. The collective has recently tested the application of the European Commission’s HLEG AI recommendations. 

AI can also reduce digital inequalities. Rich and relevant analysis and solutions powered by AI can contribute positively to the development of collective intelligence and inclusivity. It can enhance our ability to interact, offer varied opinions and respect diversity while increasing the sources of information or knowledge. 

We take part in numerous forums and think tanks working along these lines: Digital Society ForumImpact AIGlobal Network Initiative (GNI) and more.

Working with Arborus, we have just launched the International Charter for Inclusive AI so that diversity is respected and the whole data value chain is based on a responsible approach that enables discriminatory biases to be identified and controlled.

For us, recruiting a broad range of talent to develop AI solutions is an important issue because it reduces stereotyping and ensures objectivity. However, since the 1980s, the number of women working in the high-tech sector has steadily declined. They remain underrepresented in disciplines such as mathematics, computer science, neuroscience etc. To limit the risks of bias in AI, beyond gender, it is imperative that development upholds diversity to ensure society is fairly represented.

 

Introducing everyone to AI

According to the 2020 Edelman Trust Barometer, 51% of Europeans are worried about their jobs due to a lack of skills, especially technical. 

We’re aware that evolving technology and the impact this has on the world require continuous adaptation.

That’s why we’re developing new training programmes to build a broader talent pool to address the different challenges related to AI and data, and help our employees learn new skills and expertise in several key areas.

 

 

Orange Campus, our training network, contributes to many initiatives: 

As a long-term provider of on-the-job training, Orange has opened an Apprentice Training Centre thereby confirming its commitment to offer everyone the keys to a responsible digital world. The aim is to teach people new skills for future jobs via new courses such as becoming a Data Analyst. 

Numerik’up, inspires 100 young people aged 15 to 25 from the Paris region to learn the skills they need to work in the tech sector. By introducing them to a range of digital professions, it inspires them to learn. The initiative has been set up by Orange and the Columbus association and runs from February 2020 to May 2021.

Microsoft AI School powered by Simplon helps address the lack of digital skills by offering people, many of whom are unemployed, qualifications in AI and data science. The company also offers a wide range of training options to help employees develop their skills in data and AI.

For young graduates, the Business & Decision subsidiary of Orange Business Services launched a Data School in 2019 to offer additional training in data engineering and data science along with the opportunity to take part in concrete customer projects.

 

 

To train experts, we finance numerous doctorates in the field in collaboration with universities. We also offer certification programmes to employees such as the Data Science Project Management course, in partnership with the Ecole Polytechnique. More than 150 Group employees have been certified since 2015.

Finally in terms of the general public and in collaboration with Objectif IA, we’re helping to teach 1% of the French population more about AI. The e-learning platform launched on 2 April 2020.

 

 

We’re also taking part in Explor’IA 2020

This national event sponsored by the French Ministry of Education & Ministry of State for the Digital Sector also aims to educate the general public about AI and is supported by the Impact AI think tank and partners. Explor’IA 2020 includes MOOCs, webinars and videos aimed at a wide audience.

Finally a number of jobs are opening up in the fields of AI and Data

 

 

 

Creating useful innovation

Data analytics, AI and automation are opening up new opportunities for people, society and the environment. This is driving force behind our innovation policy, which focuses on how to better manage cities or regions, transport, health, resources (energy, wealth, etc) and knowledge.

 

AI: positive impact on smart city management 

AI enables real-time data to be collected and analysed to improve the way a city manages energy, waste or water consumption. It delivers information on traffic jams, population flow and weather patterns. In the future, AI could be used in association with augmented or virtual reality to stimulate and automate urban planning, construction law, flood plain avoidance and more.

AI and data: enabling medical progress

From diagnostics to computer-assisted surgery, companion robots, predictive medicine, analysing medical images and more, AI is central to future developments in healthcare. It can help to diagnose and treat diseases more efficiently and also speed up the invention of new vaccines or treatments.

Data analysis is also useful when it comes to monitoring the spread of epidemics. For example, during the Covid-19 crisis we have been providing anonymised data from our mobile networks to health and government authorities in several European and African countries. By helping them track population movements they can anticipate how the epidemic might evolve and adapt their healthcare resources accordingly.

 

AI: a helping hand in agriculture and food production

Agriculture is enhanced by AI through tools such as automated data collection and corrective action to detect and combat diseases and other problems that can affect crop yields. 

Sobloo, available via Orange Business Services, provides Earth Observation (EO) data from Copernicus to better:

  • Apply regulation imposed by environmental authorities
     
  • Increase crop yields to feed increasing populations
     
  • Use the latest AI technologies to estimate production needs 

In Senegal, Orange researchers are working with mango farmers on AI techniques that can detect ripening and improve harvests. 

AI: efficient industrial processes. The port of Antwerp, the second largest port in Europe, uses a large-scale data analysis solution provided by Business & Decision to monitor shipping levels and optimise operational management. AI also promises to help manufacturers simplify their value chains and adapt to a constantly changing environment.

 

 

Mobile banking will become more efficient too, thanks to Djingo, the AI virtual assistant

Orange Bank has introduced a new relationship model for connecting customers to their banking services. Thanks to the virtual advisor Djingo, they can access support via chat 24/7. Djingo understands almost 80% of the questions asked in natural language. It can also carry out simple transactions.

The potential for AI and data is significant, and we plan to advance in step with our customers, partners and stakeholders and in accordance with the principles of ethics and responsibility set out with European bodies. This also forms the basis for our corporate philosophy and purpose.  

 

Find out more