Telecom Social Partners' Joint Declaration on AI

Artificial Intelligence: a blueprint for unions and telecom companies

Technology and digital innovation have drastically transformed work and employment in unprecedented ways. Though they are not new as such, AI technologies have gained new momentum and, together with big data and robotics, they have rapidly evolved and changed the way we work. These technologies are expected to significantly impact the labour market of the future, particularly as traditional jobs and tasks transform or disappear, and new ones emerge.

The European social partners in the telecom sector, UNI Europa ICTS and ETNO, acknowledge the beneficial potential of Artificial Intelligence (AI) for innovation, e.g. in the area of medicine, mobility or communication.

In the employment context, AI systems could take over repetitive or dangerous jobs, enabling workers to focus on more significant or creative tasks. More broadly, the implementation of AI systems should ideally improve the citizen’s life and benefit not only people, but also the planet by driving solutions to combat climate change.

Nevertheless, with new technologies come both opportunities and challenges. In the case of AI, the challenges we must tackle are related to skills and training, health and safety, ethics, privacy and data protection, equality and fundamental rights[1].

We appreciate the work already undertaken at European level, especially regarding the development of ethical and trustworthy Artificial Intelligence. The Ethics Guidelines for Trustworthy AI and the Policy and Investment Recommendations for Trustworthy AI, from the European Commission’s High-Level Expert Group on Artificial Intelligence[2] as well as the Commission’s White Paper on AI[3] are particularly valuable contributions.

With this Joint Declaration, the European social partners in the telecom sector wish to contribute to the general debate on AI and to provide guidance to the sector stakeholders regarding the opportunities and the risks that AI represents.

We favour a humans-in-command approach to AI, meaning that humans should remain in control of all technology[4]. We also firmly support respect for human rights as a cornerstone value in the use of all AI technology. AI and other emerging technologies should not hinder individual well-being and help build a sustainable and inclusive society.

As social partners, our responsibility is to help shape how AI systems are designed, developed, and regulated, and to raise awareness about ethical concerns among AI developers and users.

Without social dialogue, we cannot build the best strategy for AI. European trade unions and employer organisations as social partners have an essential role to play in resolving complex questions regarding employment, training, the nature of work, inequality, and social protection systems. Decision-makers at the national and EU levels must listen to and cooperate with both unions and employers. And we must all recognise and use the expertise of those who are immediately involved in, or affected by, the design and use of AI systems. On a practical level, social partners should take part in establishing appropriate norms and implementation mechanisms, developing and monitoring training, and governing accountability.

This Joint Declaration is our blueprint for the social partners in the telecom sector to roll-out a flexible framework for ethical AI across countries and across all levels (European, national and local). It provides the key principles agreed upon by both Trade Unions and telecom companies, with a view of creating the basis for the use of AI to the benefit of both societal and economic objectives.

 

1. An ethical approach to Artificial Intelligence

The design and implementation of AI systems is closely linked to ethical concerns about the “nature” of AI and how it impacts e.g. on data-based decision-making processes or human-machine interaction. In the international debate on AI, this has led to the concept of “trustworthy AI” promoted by the OECD’s recommendations on the subject[5] and the European Commission High-Level Expert Group approach.[6]  We also firmly support respect for human rights as a cornerstone value in the use of all AI technology.

For UNI Europa ICTS and ETNO, an ethical approach to trustworthy AI is especially important when we look at the way AI systems are designed and implemented at the workplace. The challenges regarding data collection and management have ethical, practical, and legal dimensions.

 The ethical framework for the design and implementation of AI systems should be established at EU level to ensure an internal and external markets level playing field. This framework should guarantee the fundamental rights of all concerned persons; especially the employees involved in the development and use of the system.

Furthermore, the use of AI in sensitive business operations can have a great impact on an organisation’s workforce: automation of human resources deserves particular attention, as the results of AI-based decisions can influence the destiny of potential and existing employees.

To ensure that not only European companies adhere to ethical principles, we call on the European institutions to take appropriate legal action to ensure that rules and laws on AI services also apply to digital AI service providers established outside the EU, who use and/or sell their products and services in the EU.

 

1.1 Data Collection & Management

The success of Artificial Intelligence depends on the availability of large, high-quality training datasets.  The collection, management and use of data that feed AI must comply with European ethical norms, legal requirements and fundamental rights that together form an ‘‘ecosystem of trust’’.

The Social Partners recall that the processing of personal data in the employment context must abide by the obligations and safeguards of the General Data Protection Regulation (GDPR).  Furthermore, European Member States have been empowered to grant further protections to employees’ personal data.  The GDPR framework fully applies to the processing of employee data to train and use Artificial Intelligence systems and to the application of AI in the workplace.

The GDPR is a principles-based law designed to be future-proof and adaptable to emerging technologies.  Its risk–based approach allows for considering the risks and harms to individuals and to calibrate compliance accordingly.  The Social Partners believe that the risk-based and technology-neutral approach of the GDPR warrants a level of data protection that is adequate to the risk of the respective processing also with regards to AI development, deployment and use in the employment context.  The GDPR offers a framework for organisations to ensure trust in and acceptance of AI by all. A high level of harmonisation and common interpretation of how GDPR provisions apply to Artificial Intelligence is the key to nurturing trusted innovation across the EU.

1.2 Fairness and Accountability of AI Decisions

Another aspect of the broader introduction of more advanced computer systems is the possibility to delegate decisions from humans to computers. While this is an opportunity for increased productivity, it presents some challenges if decisions taken by AI systems are biased or incorrect.

The Social Partners stress that transparency is crucial to build trust in new technologies.  AI systems should be as transparent as possible for employees, to let them understand both how their data is used, and how decisions affecting them are taken.  The degree of insights required should depend on the technical feasibility and the complexity of the system as well as on its impact on individual rights.

Algorithmic bias and resulting employment discrimination are particularly concerning phenomena that need addressing to ensure that technology works in the service of people, not against them.  Transparency alone is not enough. Companies should provide for robust mechanisms that mitigate the unwanted effects of AI-based decisions and that help employees to ensure the negative impact of AI on their rights is averted or corrected.

Whereas it could be challenging to prove the absence of bias in the datasets at input level, it is certainly critical that employers avoid bias in the result of AI decisions at output level. A key principle to prevent poor or even dangerous decisions is ensuring that humans have oversight over AI systems and are ultimately responsible for their decision.

The implementation of AI systems raises many concerns and fears as regards the nature of the data that is collected and what it is used for (for instance, the use of AI for biometrics, facial recognition etc. in HR processes or monitoring tools).  

We think that it is necessary to recognise the role of social partners to raise awareness about these challenges and to discuss ethical boundaries of AI systems and the data collection/management they require. This could be for example by integrating the discussion on AI governance into existing social dialogue structures (e.g. works councils…) and/or create advisory boards (with the participation of social partners).

In order to make this line of responsibility function well, we need to ensure that employers establish adequate mechanisms for data governance and accountability reflecting ethical considerations, and that the social partners are fully involved in them. Human oversight models should be proportional to the risks involved by the AI application at hand.

Decisions made by algorithms can lead to less democratic workplaces if they are not followed up with adequate supervision. Therefore, social partners should focus on essential principles for avoiding bias, including, but not limited to: transparency, traceability and neutrality of AI systems (in terms of age, gender, ethnicity, and political standing).

Trustworthy Artificial Intelligence must not be an empty slogan. The Social Partners remain committed to playing their part in bringing trust to the application of AI in the workforce and to reconciling innovation, efficiency, and protection of employees’ data and fundamental rights.

 

2. The need of a building site of European AI skills

As the European Union and its Member States approach the question of how to articulate recovery policy and spending, we believe that there is a major building-site to be priorities across countries: digital skills and training.

Artificial Intelligence, as well as any type of constant change in technology, leads to major transformations with direct impact on skills and training. This is particularly true in the telecommunication industry, which has faced significant disruption in the past years, in the wake of the smartphone and app revolution. At the same time, as we look ahead, major telecoms-lead innovations like 5G and its applications are expected to create a wealth of opportunities for the telecom ecosystem at large. The main underlying question is how effectively and how fast we can train the necessary skills to ensure that all workers involved in the sector, as well as partners in the value chain and clients, keep up with both, past disruption and future opportunity. For this reason, the number one priority is to boost hard and soft skills associated with ongoing and new and emergent technologies in the sector, while ensuring adequate resources are dedicated to training and re-training.

This effort cannot be just the effort of the telecom industry and of its workers. Because of the high value-added that telecom innovation can bring to society, it is in the primary interest of all stakeholders – public and private – to prioritise and scale up digital skilling and training programmes.

The standard modus operandi should foresee the engagement of all those who are deemed to be impacted by, or benefit from, technological change: industry, government, social partners, schools, colleges, universities, training providers are essential actors in  supporting this transformation process and ensure nobody is left behind. As European Governments move to develop digital skills programme and prioritise technology in the use of the EU recovery funds, they should enable both interaction across industries and sectors of society, and funding digital upskilling programmes within the telecom sectors, in its value-chain and with SMEs.

As of today, the telecoms sector has identified two key challenges: on the one hand, attracting and recruiting  new and competent talent from a much broader and diverse  group: on the other hand, the necessity of further training and upskilling the existing workforce. Developing initiatives for the digital upskilling and for promoting diversity within the European telecoms sector, with a particular focus on gender equality, ageing workforce and vulnerable groups, will be a key to success.

Furthermore, a constant update of skills will be more and more important in the future and this should be done in close co-operation between employer and employees. This calls for a constant discussion and a trust-based cooperation between social partners at all levels (European, national, industry and company). As an example, a common skill set that might be required includes such skills as: creative ways to solve problems critical thinking, constant learning, decision making, interaction and co-operation skills, empathy, systemic thinking, being able to withstand uncertainty and pressure and intuition.

In short, the following areas of focus should constitute the lighthouse for all AI policies in the sector:

  1. Within the telecoms industry:

    • Re and up-skilling of current workforce
    • Hiring of diverse talent to handle AI
  2. In the tech and telecoms ecosystems:

    • Education of partners along the value-chain
    • Activate stakeholders and partners on ethical matters
    • Facilitate uptake by/awareness among customers
  3. Through policy and government action:

    • Support to industry schemes for re and up-skilling of current workforce
    • Boost education and training programmes for AI and AI ethics
    • Promote AI education and uptake across sectors, with special regard to SMEs and Public Administration

The Social Partners remain committed to advancing the telecom industry’s overarching objectives. Most notably, UNI Europa ICTS and ETNO are actively supporting the training and upskilling of the ICT workforce in order to bridge the digital skills gap and have been members of the Digital Skills and Jobs Coalition[7] since February 2018.  Furthermore, we are engaged in a social dialogue project that aims to develop best practices and create recommended pathways for the digital upskilling and inclusion of the diversity group within the workforce of the European Telecoms sector[8].

Europe has the opportunity to activate some of its key assets and lead in the development and uptake of human-centric AI. As the pandemic dramatically forces an acceleration of tech uptake, EU institutions and Governments should seize the moment and ensure AI awareness, education, investment and innovation is scaled up across territories and across industrial sectors. This must be accompanied by an ethical implementation of AI through savvy public policies reflecting the concerns of Europe’s citizens.  The European social partners in the telecom sector are committed to play their part in this process.

 

[1] Cf. ILO Declaration on Fundamental Principles and Rights at Work, https://www.ilo.org/declaration/thedeclaration/textdeclaration/lang--en/index.htm

[2] High-Level Expert Group on Artificial Intelligence, European Commission (April 2019): Ethics Guidelines for Trustworthy AI. Brussels: European Commission. https://ec.europa.eu/futurium/en/ai-alliance-consultation/guidelines and AI HLEG, EU (June 2019): Policy and Investment Recommendations for Trustworthy Artificial Intelligence. Brussels: European Commission. https://ec.europa.eu/digital-single-market/en/news/policy-and-investment-recommendations-trustworthy-artificial-intelligence .

[3] European Commission (February 2020): White Paper On Artificial Intelligence - A European approach to excellence and trust. Brussels: European Commission. https://ec.europa.eu/info/sites/info/files/commission-white-paper-artificial-intelligence-feb2020_en.pdf

[4] This approach has also been adopted by the European cross-sectoral social partners in their recent Framework Agreement on Digitalisation, https://www.etuc.org/system/files/document/file2020-06/Final%2022%2006%2020_Agreement%20on%20Digitalisation%202020.pdf

[5] Council on Artificial Intelligence, OECD (2019).

[6] AI HLEG, EU (June 2019).

[7] https://ec.europa.eu/digital-single-market/en/digital-skills-jobs-coali…

[8] Link to the full press release: https://www.etno.eu/news/all-news/672:uni-europa-etno-dufa.html