Commission’s AI policy: some progress but still too little to protect workers

Thursday 20 February 2020

Editorial by UNI Europa Regional Secretary Oliver Roethig.

Yesterday 19 February 2020, the European Commission has published its long-awaited White Paper on Artificial Intelligence (AI).

UNI Europa welcomes the White Paper as a further step towards a European AI strategy that considers both the opportunities and risks of AI and should be based on the respect of ethical concerns, European values and human rights.

Although the Commission refers to the involvement of social partners in general, the text does not reflect the importance of social dialogue at all levels of AI design and deployment. Without social dialogue, we cannot build the best strategy for AI. In our recent position paper on AI, UNI Europa has highlighted the key role European trade unions and employer organisations must play in resolving complex questions regarding employment, training, the nature of work, inequality, and social protection systems.

The involvement of social partners is especially important with regard to data and AI governance. The Commission’s White Paper promotes the creation of more data and even refers to large volumes of data that are “under-used”. We are concerned about this somewhat naïve belief that more data alone will be enough to boost the European economy. On the contrary, we are convinced that a solid European AI strategy must provide concrete governance mechanisms to ensure data protection and the respect of privacy rights.

The Commission has correctly analysed the risks linked to possible flaws in the AI system design or data use. This is why, complementary to regulation, social dialogue is essential as trade unions and companies should create governance mechanisms through which data collection can be monitored, regularly evaluated, and ameliorated.

We appreciate the Commission’s acknowledgement that we need to upskill the workforce and should in particular pay attention to attract more women into the ICT sector. Likewise, the introduction of ethical guidelines into an “indicative curriculum” for AI developers echoes our demands put forward in the UNI Europa position paper. Unfortunately, the Commission fails to notice that training and upskilling efforts should not be limited to high-level technical skills alone. In fact, automation and AI system deployment actually increases the importance of soft skills, including creativity, empathy, and complex reasoning.

In conclusion, UNI Europa welcomes many of the White Paper’s valuable analyses and proposals, but we still think that there is room for improvement.