The European Parliament is seeking a common EU approach to
the development of automated decision-making processes in order to reap the
benefits of these processes while avoiding risks and the fragmentation of the internal market
in terms of regulation. In its Resolution on ‘Automated
decision-making processes: Ensuring consumer
protection, and free
movement of goods and services’,
MEPs have called for consumer
rights to be strengthened and
to ensure that final and permanent automated decisions are subject to
The European Parliament welcomes the potential of automated decision-making, but believes that when consumers interact with such a system, they must be properly informed about how the system works,
how to contact a person with decision-making powers, and how to review and correct decisions made by the system.
with the recommendations
put forward by the Commission’s Expert Group in
its Ethics Guidelines for Trustworthy Artificial Intelligence (AI),
Parliament is thus placing emphasis on human control and, where appropriate,
final decision-making. The Commission is called upon to examine whether
regulatory gaps exist and whether measures are needed beyond existing
regulations to protect consumer rights in connection with AI and automated
Automated decision-making in social security
The issue of trust in automated decisions and their
monitoring could also become increasingly important in future processing by social
security institutions. For example, since autumn 2019 the Swiss accident
insurance system has
been testing partially automated processes to identify claims that meet the
definition of an accident under Swiss law.
In addition, automated decisions will also have an impact on
the safety of machines and human-machine interactions, and thus on prevention issues in the workplace.
Safety when using machines
The European Parliament points out that products with
automated decision-making capabilities can evolve and ‘act’ in ways that were not
foreseen when they were first placed on the market.
The Commission is called upon to put forward proposals to adapt safety
rules for a number of specific European product safety regulations, such as the
Machinery Directive, as well as for ‘non-harmonised products’ covered by the General Product Safety Directive. MEPs want to
protect users and consumers from harm and provide manufacturers with clarity about their obligations.
The European Parliament considers that a risk-based
regulatory approach is necessary, given the complexity of the various
applications of AI and automated decision-making systems. In a report commissioned by the German government, the German Data Ethics Commission also recommends a risk-adapted
regulatory approach for algorithmic systems. According to the Data Ethics Commission, this should
be based on the principle that the
greater the potential for harm, the more stringent requirements and
intervention by regulatory instruments must be.
In assessing the potential for
harm, the entire socio-technical system must be taken into consideration, that is, all
components of an algorithmic application including all people involved, from the development
phase (e.g. with regard to the training data used) through to its implementation in an application
environment and its assessment and any necessary
MEPs are calling on the Commission to develop a
risk assessment system for AI and automated decision-making
to ensure a consistent approach to the enforcement of product safety
rules in the internal market.
As announced in its Work
Programme, on 19 February 2020 the Commission presented a White Paper on developing and utilising AI, while also
ensuring full respect for European values and fundamental rights. AI
could help to find new solutions to old problems and speed up a wide range of different work
processes. However, this requires an ecosystem of trust to ensure
it develops within clearly defined ethical boundaries. At the same time, it launched a public consultation on the White Paper. Interested parties can participate until 19 May 2020.