Print
See related documents
Committee Opinion | Doc. 14303 | 26 April 2017
Technological convergence, artificial intelligence and human rights
Committee on Legal Affairs and Human Rights
A. Conclusions of the committee
(open)1. The Committee on Legal Affairs
and Human Rights congratulates the rapporteur of the Committee on Culture,
Science, Education and Media, Mr Jean-Yves Le Déaut (France, SOC),
on his report, and supports by and large the proposed recommendation.
2. The committee proposes some amendments aimed at highlighting
certain legal aspects of the recommendation. There is no doubt that,
given the complexity and the rapidly evolving nature of emerging technologies,
many more legal and human rights aspects ought to be thoroughly
considered. The Parliamentary Assembly should continue looking into
this issue in the not too distant future.
3. Furthermore, as rightly pointed out in the draft recommendation
tabled by the Committee on Culture, Science, Education and Media,
the committee underlines the pressing need for the Committee of
Ministers to conclude its work on the modernisation of the Convention
for the Protection of Individuals with regard to Automatic Processing
of Personal Data (ETS No. 108) by adopting the revised convention
as soon as possible.
B. Proposed amendments
(open)Amendment A (to the draft recommendation)
After paragraph 3, insert the following paragraph:
“The Assembly recalls the principle enshrined in Article 2 of the Convention on Human Rights and Biomedicine (ETS No. 164, ‘Oviedo Convention’) which affirms the primacy of the human being by stating that ‘the interests and welfare of the human being shall prevail over the sole interest of society or science’.”
Amendment B (to the draft recommendation)
In paragraph 6.1, after the word “finalise”, insert the words “without further delay”.
Amendment C (to the draft recommendation)
Before paragraph 8.1.1, insert the following paragraph:
“the fact that responsibility and accountability lies with the human being, no matter the circumstances. References to ‘independent’ decision making by artificial intelligence systems cannot exempt the creators, owners and managers of these systems from accountability for human rights violations committed with the use of these systems, even in cases where the action that caused the damage was not directly ordered by a responsible human commander or operator;”
Amendment D (to the draft recommendation)
After paragraph 8, insert the following paragraph:
“The Assembly reiterates its call made in Resolution 2051 (2015) “Drones and targeted killings: the need to uphold human rights and international law” on all member States and observer States, as well as States whose parliaments have observer status with the Assembly, to refrain from any automated (robotic) procedures for selecting individuals for targeted killings or any sort of injury based on communication patterns or other data collected through mass surveillance techniques. This should be true not only for drones but also for other combat equipment with artificial intelligence systems, as well as other equipment and/or software which might potentially inflict damage on people, property, personal data or information databases, or interfere with privacy, freedom of expression, or the right to equality and non-discrimination.”
C. Explanatory memorandum by Mr Boriss Cilevičs, rapporteur for opinion
(open)1. I would like to congratulate
Mr Le Déaut on his report, which rightly highlights the challenges
posed to human rights by the emerging technologies.
2. Let me stress from the outset that it is not easy to provide
a legal point of view on the questions raised in the report. In
fact, we lack the expertise to do so, since very little, if any,
such expertise is available at present. What is described in the
report may sound like science fiction. The technological (r)evolutions
are of such a scale that they are overwhelming. Nevertheless, they
are very real and it is of utmost importance that we seriously consider
their possible consequences on human rights.
3. The peculiar feature of the “genuine” artificial intelligence
systems that makes them different from “common” software is their
ability to produce results not directly envisaged by instructions
set by the programmer. This is done through “independent” accumulation,
processing and taking account of new data. Therefore, artificial
intelligence systems, in some sense, possess the ability to learn
and take decisions on their own.
4. The subject matter is far too vast and complex to be adequately
dealt with from a legal perspective in an opinion. I will therefore
limit myself to a few comments. I suggest that we continue working
on this issue more in depth in a future report.
1. Accountability issues
5. From the legal point of view,
this report raises the issue of the subjectivity of the artificial
intelligence system. The issue comprises two aspects. First, may
it be considered an independent right-holder? This issue has been
broadly discussed in science fiction, and so far remains an exclusive
competence of the latter. However, another aspect, that of accountability,
has already become topical in real life. Who bears the responsibility
for damage inflicted by the artificial intelligence system, including
violations of human rights?
6. The Parliamentary Assembly has already partly explored this
issue. In particular, in its Resolution
2051 (2015) “Drones and targeted killings: the need to uphold human
rights and international law”, the Assembly called on all member
and observer States, as well as States whose parliaments have observer
status with the Assembly, to refrain from “any automated (robotic)
procedures for targeting individuals based on communication patterns
or other data collected through mass surveillance techniques”.
7. In its recent resolution on Civil Law Rules on Robotics, 
the European Parliament stresses
that draft legislation is urgently needed to clarify liability issues,
especially for self-driving cars. In this resolution, the European
Parliament rightly states that, in such a context, “the legal responsibility
arising from a robot’s harmful action becomes a crucial issue”.
The European Parliament goes as far as asking the European Commission to
“consider creating a specific legal status for robots in the long
run, in order to establish who is liable if they cause damage”.
the European Parliament stresses
that draft legislation is urgently needed to clarify liability issues,
especially for self-driving cars. In this resolution, the European
Parliament rightly states that, in such a context, “the legal responsibility
arising from a robot’s harmful action becomes a crucial issue”.
The European Parliament goes as far as asking the European Commission to
“consider creating a specific legal status for robots in the long
run, in order to establish who is liable if they cause damage”.8. In my view, however, it should be clear that accountability
lies with the human person, no matter the circumstances. In any
case, references to “independent” decision making by artificial
intelligence systems cannot exempt the creators, owners and managers
of these systems from accountability for human rights violations
committed with the use of these systems, even in cases where the
action that caused the damage was not directly ordered by a responsible
human commander or operator.
2. Relevant Council of Europe instruments
9. Two Council of Europe conventions
are of direct relevance to the issues at stake.
- The Convention for the Protection of Human Rights and Dignity of the Human Being with regard to the Application of Biology and Medicine: Convention on Human Rights and Biomedicine – “Convention on Human Rights and Biomedicine” or “Oviedo Convention” (ETS No. 164 )
10. The Oviedo Convention protects human rights in the biomedical
field. It is a framework convention aimed at protecting the dignity
and identity of all human beings and guaranteeing everyone, without
discrimination, respect for their integrity and other rights and
fundamental freedoms with regard to the application of biology and
medicine.
11. Very importantly, the Oviedo Convention affirms the primacy
of the human being. Article 2 of the Oviedo Convention reads as
follows: “The interests and welfare of the human being shall prevail
over the sole interest of society or science.” This is especially
relevant in the field of care robots. As pointed out in Mr Le Déaut’s report,
this raises the question of the right to choose between human contact
and assistance by a robot. Ultimately, if human contact were to
be totally replaced – with the effect of potentially leading to
complete deprivation of human contact – Article 3 of the European
Convention on Human Rights (ETS No. 5) might be at stake.
- The Convention for the protection of Individuals with regard to Automatic Processing of Personal Data (ETS No. 108)
12. Convention No. 108, opened for signature on 28 January 1981,
was the first – and is today still the only – legally binding international
instrument in the data protection field. Under this convention,
the parties are required to take the necessary steps in their domestic
legislation to apply the principles laid down in order to ensure
respect in their territory for the fundamental human rights of all
individuals with regard to the processing of personal data. Obviously
there have been tremendous evolutions since 1981 in the field of
new information and communication technologies, especially when
it comes to the automatic processing of data. These evolutions bring
new challenges for privacy.
13. Modernisation has therefore become necessary in order for
Convention No. 108 to reflect new realities and to continue to provide
adequate protection in the light of evolving and emerging technologies.
In 2010, it was decided to undertake such a modernisation of the
convention. The aim is to better address the challenges resulting
from the use of new information and communication technologies,
but also to strengthen the implementation of Convention No. 108.
14. The relevant intergovernmental committee (ad hoc committee
on data protection – CAHDATA) finalised its work in June 2016. At
that time, the Secretary General of the Council of Europe expressed
the hope that the work would be concluded at the level of the Committee
of Ministers by the end of 2016. However, to date, the revised treaty
has still not been adopted.
However, to date, the revised treaty
has still not been adopted.15. The Assembly should therefore strongly urge the Committee
of Ministers to conclude its work on the modernisation of Convention
No. 108 and to adopt the revised treaty without delay.
3. Amendments
Amendment A is self-explanatory.
Amendment B is self-explanatory.
Amendment C
Explanatory Note:
This amendment aims at clearly stating that accountability, under any circumstances, lies with the human being.
Amendment D
Explanatory Note:
This amendment aims at reiterating the call made in Assembly Resolution 2051 (2015), as mentioned above, while widening this call by not only addressing drones (which were the specific topic of the 2015 report) but also other artificial intelligence systems which might potentially inflict damage and breach human rights.