Judicial Integrity, IT and AI in Doha, Qatar
On 25 and 26 February 2020, the Global Judicial Integrity Network held its second High Level
Meeting, in Doha, Qatar. The overarching topic is always increasing judicial integrity,
but this time the two big themes were gender diversity and information
technology. My contribution consisted of a plenary panel discussion on
information technology and artificial intelligence. Here is a summary of my
contribution to the discussion.
Left
to right: prof. Karen Yeung University of Birmingham, judge Madan Lokur from
India, panel chair Roberta Solis, Judicial Integrity team leader UNODC, myself, Judge Victor
Momotov of the Russian Federation, Judge Ju Yeon Lee, from the Republic of
Korea. (thanks to Judge Kees Sterk for the picture)
Roberta Solis sent
us all questions beforehand.
My first question
was on experience with in house IT development in the Netherlands.
Like most other
judiciaries, the NL judiciary uses many different kinds of information
technology.
Case management,
office technology, multiple web sites with information, news and published case
law (50.000+ judgments annually), court management tools, an intranet, email,
and some e-filing and digital procedures. All this is managed by the judiciary’s
own IT organization, under the Council for the Judiciary of the Netherlands.
An important reason
for choosing in house development is continuity in knowledge: keeping the
knowledge that is developed inside the judiciary. But there is a more important
reason that is not always recognized. IT has reached a level of development
that includes all work processes. It is no longer a tool, it is becoming an
environment. And since it includes the primary, judicial work processes, the
choice was to do that under judicial control.
The way to achieve
this was to mix in house expertise with individually contracted specialist experts.
The methodology of choice was agile product development, with judges as product
owners. This ensured strong judicial involvement in the software development.
Some lessons to
take away from this experience so far:
Development is not
the most difficult issue. The most difficult issue is governance: changing work
processes and implementing digital procedures require strong decision making
processes geared for innovation. Most judiciaries are geared to be production
organizations: they process cases. For going digital, this governance setup
needs to change.
There are also some
options for innovation to be gleaned from the experience in the Netherlands. The
most important variables are the need for changes in the regulation and the length
of the procedure to be digitalized.
1 simple, short
term processes can be replaced in one operation if they do not require changing
regulations
2 more complex,
longer running processes should be changed step by step, especially if they
also require changes in regulation
3 setting up an entirely
new process avoids the risks involved in option 2
4 setting up an
entirely new institution creates the possibility to design approaches that
leverage the power of going digital.
Roberta’s second
question for me was about the CEPEJ charter for ethical use of artificial
intelligence (AI) in the administration
of justice.
CEPEJ is the
Council of Europe’s Commission for the Efficiency of Justice. CEPEJ decided to draw
up the Ethical Principles in order to ensure responsible use of artificial
intelligence in compliance with the European Convention on Human Rights and the
Convention on the Protection of Personal Data, to help improve the
predictability of the application of the law and the consistency of court
decisions and to prevent discrimination.
In June 2018, CEPEJ held a session with
experts on the workings of courts, legal experts and experts on the anthropological
and philosophical aspects of the emergence
of artificial intelligence. CEPEJ also conducted a survey of the state of use of
AI in the judiciaries, which showed that that use is still very low.
Experts produced
reports on operating characteristics, legal reasoning, explaining judges’
behavior in retrospect, data protection, and predictive justice tools, that are
included in the final reports as annexes.
The 5 principles in
the final document are:
1
Respect
for fundamental human rights
2
Non-discrimination
3
Quality
and security
4
Transparency,
impartiality and fairness
5
“under
user control”
The Ethical
Principles, therefore, also pose some new challenges for judiciaries.
The Principles need
to be operationalized, in regulation as well as organization. This is going to
be a big challenge for judiciaries, as they will need to safeguard the
integrity of their own data and their information systems. They also need to make
their information machine readable in order for AI to make more effective use
of it.