Automated profiling in law enforcement? What can we know about the actual practice in Germany?

Video in TIB AV-Portal: Automated profiling in law enforcement? What can we know about the actual practice in Germany?

Formal Metadata

Automated profiling in law enforcement? What can we know about the actual practice in Germany?
Title of Series
Number of Parts
CC Attribution - ShareAlike 3.0 Germany:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this license.
Release Date

Content Metadata

Subject Area
Profiling technologies have impact on societal norms and practices. A research team at Technische Universität Berlin conducted a case study on the role of profiling technologies in the field of law enforcement and monitoring of political activism. In this talk, first outcomes will be presented.
Presentation of a group Decision theory View (database) Multiplication sign Archaeological field survey Data analysis Disk read-and-write head Information privacy Tracing (software) Computer programming Neuroinformatik Sign (mathematics) Data acquisition Information Process (computing) Fundamental theorem of algebra Covering space Area Algorithm Block (periodic table) Linear regression Constructor (object-oriented programming) Lattice (order) Data mining Process (computing) System identification Right angle Figurate number Fundamental theorem of algebra Slide rule Game controller Observational study Rule of inference Product (business) Number Broadcasting (networking) Frequency Cross-correlation Profil (magazine) Term (mathematics) Codierung <Programmierung> Associative property Informationsgesellschaft Form (programming) Focus (optics) Information Projective plane Computer program Mathematical analysis Database Basis <Mathematik> Basis <Mathematik> Personal digital assistant
Group action Presentation of a group Multiplication sign Modal logic Source code Open set Tracing (software) Perspective (visual) Neuroinformatik Blog Different (Kate Ryan album) Hypermedia Core dump Software framework Information Kolmogorov complexity Physical system Fundamental theorem of algebra Algorithm Theory of relativity Block (periodic table) Feedback Constructor (object-oriented programming) Keyboard shortcut Moment (mathematics) Interior (topology) Data storage device Parameter (computer programming) Staff (military) Instance (computer science) Connected space Degree (graph theory) Process (computing) Symmetry (physics) Order (biology) Self-organization Right angle Quicksort Arithmetic progression Fundamental theorem of algebra Internationalization and localization Point (geometry) Game controller Link (knot theory) Computer file Observational study Civil engineering Line (geometry) Virtual machine Digital library Content (media) Focus (optics) Event horizon Field (computer science) Power (physics) Frequency Broadcasting (networking) Profil (magazine) Authorization Energy level Boundary value problem Data structure Task (computing) Form (programming) Electronic data processing Addition Focus (optics) Information Validity (statistics) Expression Physical law Computer program Mathematical analysis Content (media) Interactive television Expert system Analytic set Core dump Database Group action Limit (category theory) System call Mathematics Uniform resource locator Event horizon Visualization (computer graphics) Personal digital assistant Enumerated type Object (grammar) Local ring
Context awareness State of matter Modal logic Multiplication sign Data analysis Parameter (computer programming) Mereology Perspective (visual) Web 2.0 Programmer (hardware) Mathematics Core dump Fundamental theorem of algebra Area Algorithm Moment (mathematics) Instance (computer science) Flow separation Demoscene Type theory Exterior algebra Telecommunication Quicksort Web page Slide rule Game controller Functional (mathematics) Service (economics) Observational study Field (computer science) Metadata Revision control Confluence (abstract rewriting) Internetworking Energy level Form (programming) Electronic data processing Focus (optics) Information Validity (statistics) Weight Projective plane Physical law Computer program Mathematical analysis Line (geometry) Transformation (genetics) Word Personal digital assistant Logic
Data acquisition
so it
and back
we're starting with non existent is very on a very short time left to be the session from and the next 1 is the well i . e at what the and he is um is research associate at the that home of uh of this would give that stuffed heads up that usually that he had he is talking about body conceived behind me and automatic profiling in law enforcement somewhat um what can we know about but the actual practice in Germany and terms so go on and I'm I'm I think the other very interesting sessions this term for the next half hour and he said that isn't that he has made 20 minutes of farm of actual actually the of the lecture and then we have enough time to have and that have more questions because then we have a little metric had 1 of an odd to have all the best OK gone OK thank you there yeah maybe we also can continue the discussion we started right now because some of the problems are related so thanks for coming to this session my name is area was introduced it running them and the title of my presentation automated profiling in law enforcement and thanks again for the period to opportunity to % some initial findings to date from our case study about the user profiling technologies in the law enforcement sector so yeah 1st of all that and some short informations about the profiling to the project is funded by the European Union's fundamental rights Justice program and the main question is what are the impacts on profiling on fundamental rights and values accordingly the acronym profiling stands for protecting citizens rights fighting fighting this profile so to answer this question we conducted a number of opportunities last year the coordinated unique made a survey to learn more about human profiling from the national European Data Protection Authorities case studies about the user profiling for border control in Italy and the use of profiling for commercial advertising in Romania that conducted by all partners and we did a case study of other profiling and enforcement whatever because we have had and for all and further into the detail about it on case study I want to shortly introduce of working definition of profiling that I will revisit some hopes and fears about big data which is important to keep in mind when we talk about profiling afterward so we show in which ways to produce data traces of which role they play for the construction of information in a way and this would be redundant to look at in sense so I will try to the goal of Cricket about the slides and finally I will discuss the findings you made so far in our case yeah so what is profiling especially in Germany the temporal findings of the youthful offender profiling which only refer to the identification of defended to the psychological profile of a criminal but we understand profiling as a partially automated process of analyzing personal or personal data aimed at producing knowledge in the form of profiles to be used as a basis for this decision making and this is related to information science and the analysis of large databases it is thus strongly linked with the terms Data Mining and Knowledge Discovery in Databases but what means knowledge it's good to keep in mind that when data analysts and big data and assess talk about knowledge they usually mean correlations and probabilities Europe put a graphic uh from 2 encoders and broadcast this book about discrimination privacy in the Information Society these figures described 3 different algorithms used for profiling pattern mining using linear regression clustering classification I don't want to go into detail about these techniques said just want to highlight that when we talk about profiling profiling we talk about this kind of knowledge production the the so then now something about Big Data hopes and fears anyway the idea of using computers and data to make the world more efficient uncontrollable is not new since the emergence of computers there have been dreams to gain better control over the rules of societies I brought here the cover page of a 1985 issue of the child signs for the people which is a more critical view but the title computing the future very well shows which means that in the big data analysts have in the past and today the 2nd picture is taken from innovation study but potential so being on big data from the Fraunhofer Institute for intellect Intelligent Analysis and Information Systems and the 3rd picture is from the block off the chairman technology fair Siebert which had a focus on big data this year 1 can doubt what it actually means of companies of 5 times more likely to make decisions faster or if 50 % say be data helps to better meet consumer demand but numbers convince and that's a big opportunity for the data so the
other side is the fear of dehumanizing the world all computer algorithms gaining control of humanity on the 1 hand it this is expressed in the thought that humans become machines or subjected to machines like in the picture on the left than on the other hand people are concerned about the abuse of technologies to control people's thoughts and behavior but whereas the data come from basically we can categorize 3 data
sources sensors like cameras and microphones are not very new but they are more and more digitize which makes them open ICT processing and computer analytics moral we have lots of technical data traces where sensors are intended to collect data blocks are technical information at the 1st place but can be used to gain knowledge about interactions behavior for instance lots about telephone calls what phones or are intended to organize the telephone system but can be used as information about locations and social relations between individuals user generated content is the benefit of replacing one-directional broadcast media like television and radio through Natural Media where any individual consuming content can give they can share information across different and this actually turns individuals into data sources so we use all some of the horrific pictures maybe um what does the traces mean in practice for of the people who just came into the room and this is sort of the the presentation before I just want to say the the the left and the picture is a visualization of social connections to with the column to the to the end of the the project and they're on the right we see that some information which can be drawn from your mobile phone connections you molar reduces said Alice attributes annotation enriched with the information about ingoing and outgoing calls and the content so for example the future to account that's a lot that can be known about you this is this your interests to social contacts and even your recent they're about a the so far we would be revisited what kind of data is generated by us and our daily interactions and what it means to analyze this information to gain something or knowledge of or now I want to present some findings of our case studies to show which role profiling place in the law-enforcement the same in the field of political and I 1st of all I need to say that I am presenting work in progress so the analysis is going on but we have some findings to think guy to share coding you feedback is appreciated after the presentation of our idea was that when you're talking about profiling data processing and data protecting their detection the most relevant here is the impact on the exertion of fundamental rights that's why we chose to focus on political activism because the freedom of expression and political participation our core values of democracy this field of research is very sensitive because the debate about the issues is of the emotional and experiences that quite subjective however we address these problems by taking into account various perspectives we did qualitative interviews police experts lawyers scholars technology producers and political activist at this point to get started fully automated profiling as introduced in the definition is not used in the police sector of that but it seems that automation plays an increasing role nevertheless we find that some interesting hints about the changing processes of reasoning and the construction of suspicion our focus is thus more on the changing practices and the role of technology in humans less on an exhaustive enumeration of police let me start with the police perspective from the police perspective and it's important that large amounts of data become accessible through the digitization of data especially complex relations and structures can be made visible but information is stored in databases and links and relations between objects places events organisations and individuals can be visualized and thus missing links can be found this is much easier than with paper folders police petitioners feeling necessity to adapt the tools and capabilities to the increasing relevance of Globally Networked prime which says justifies the need to exchange information different apartments different authority levels which is much more efficient once information in order to but the interchangeability of data is limited by law and ability of databases is at the moment a it both limitations of based on the principle of binding but there are discussions ongoing to the specific remarkably assessing the data is of increasing relevance validity of data needs to be thoroughly assessed where does it come from what does it mean also upon his death police staff is specially trained for this task this leads to high documentation requirements it's not only important in which data is available but also how it was generated and if it has been stored correctly so it seems that the transition from traditional folders to digital databases is still in progress some interviewees reported burden of additional documentation others reported difficulties with document documentation processes another aspect is that the validity of digital information needs to be made to comprehensively in court by humans would consent and difficult to reproduce investigation and the trial is a long period of time so we see the new demands of assessing validating adoption maintains documenting information shows that the new technologies used by the police have raised uncertainties regarding the quality of the a out and finally it was repeatedly stressed by police interview is that there's no data collected or stored without legitimate grounds that means the benefits justify an initial suspicion moreover it was stressed that there is in general no interest in doing more of investigational analysis the necessary and also not enough resources and we got a form power from the lawyer's perspective delegitimated grounds for data storage and processing and sometimes awakened arbitrary notably it was also reported that lead the framework of the new methods of data collection and investigations so sometimes unclear but nevertheless evidence is generated and used during files they also stated that sometimes the police documentation of the sources of their knowledge is incomplete which makes it difficult for them to comprehend the investigation processes and the construction of suspicion has resigned lawyers you know which a symmetry in favor of the as an example it seems that it has become a common practice to do choose location varies what he saw this picture with from I it's with yeah localization of mobile phones and and the practice of these location varies goes like beyond the scope of using it only in very severe cases so often use a kind of uncertainty perceived but in this perspective it is linked to the opacity of data processing activities so let's come to the citizen perspective the opaque is also in which information is stored by the police about individuals and it can be very difficult to get those kinds of information if there is an and try and of the degrees and which data is stored and for how long but of course if you behaving it is very likely that there's nothing stored in the 1st place this uncertainty that the boundaries between behaving normally maybe suspicious was described by activists it was reported that activists withdraw from being politically active politically active and from cooperating with civil rights groups which are considered to be potentially under scrutiny thus people refrain from exerting their fundamental rights due to the uncertainty what information could be collected about them and how this could be used against them this in a way reminds of the book the prior from
unstaffed coverage very well describes how it might be not to know what information is processed about a person and which part of it is considered as suspicious this picture from the film version from 1993 shows the scene that shows of k visits the called room 1 free day it very well illustrates the subject under suspicion and which has no idea what is the core of the trial what do they know what is conceived as wrong behavior Joseph year to see is that there is a trial going on but at the moment cannot see anything we'll get any information about what is going on there it might be individual cases that people strongly experience unjustified surveillance and control but it is important to keep in mind these cases when talking about new regulation and technologies like profiling you as a summary the role of humans in data processing has become increasingly to document and assess and validate information reasoning changes weight of information changes depending on which information can be retrieved from data accessible the grounds for data collection become blurry social networks become more relevant in the future of our case study reasoning is inscribed in data analysis tools this also leads to new forms of suspiciousness the amount of data and new ways of analysis the to uncertainty the a for the nation and documentation uncertainties also on the side of citizen and lawyers because they experience difficulties in reconstructing the line of argumentation due to the opacity of data processing and information construct the that's from my side and thanks for listening and and uh very curious for discussion on and if you want further information you can visit our
project page there are also working papers and um there will also be something like this in the future In the next month when the finish of the studies and have by the rewards and thinking and on then a question and if you the you have yeah hide more monitors should but my name and I've just a little low if you can say something about the and necessity pitch types and which types of data are actually more important for profiling does just like the metadata from telecommunication and Internet usage origins of Moses validated data from insurance companies from your uh P our data from slide into trouble companies and so on like what is the most important type of data for profiling work or just like to to make it more and more concrete didn't more necessary to have the best and the vast amount of data or the more about the quality of data um that's a good question and you and what I have to say we don't have that too much information from the police and that's what really is the most important data because it also depends on on the cases and um I would say it's really strongly depends from case to case which information gathered together and put on the other hand when we come to the to the field of political activism what we have reported also from from the from lawyers and political activists and this she is and this is a very effective tool because you can easily check which persons have been known on several places for instance a 10 so I would say and that's a very important resource but yet we don't have uh part of facts about and and to and the thank you for bringing this topic I'm wondering what your experience with in all your the you know the algorithm on the state so to say because the alibi that's going to be much easier if we knew the logic behind for finding and of course there's always secured you're given that they cannot say that but I believe that if we keep it like a wide will always lose so we have to try to to find some middle ground where a certain aspects of this logic have to be revealed because there had the 2 should be there should be borders of legitimacy every not every sort of profiling with legitimate for example clearly discriminatory are taking to counter sensitive data even if the police shouldn't really be out right so did you have any ideas how to juggle these are hard to find a middle ground between knowing nothing and demanding everything which we will never get from I I would say in general we have some some I we we can we can get some information what's what other the main ideas or what what other practices for instance them if you don't the 6 they tell us that our are uh this economic crimes is a very important field and organized crime and that they they they tried to to find those links and on the other hand political activism as as the it's a stronger focus than than than other things around so I would say at the moment um it's quite focused on on on several fields and there's no no general suspicion so but on the other hand and at the same time we cannot really say where's the border and and I think uh when you say it's not like a wide and then let it slide but some interesting are the blurry parts and and it's it's it's it's hard to to find out where where where where's the border between a and B suspicious announces suspicious the the yeah I think the question I will give the 1st line segment is you know what and and mean yeah last come the the lawyer's perspective on this is always the the regulations are too unclear but if you have a changing technological context I don't know how you make clear exact regulations for every context and what what what came out and how would you suggest in that regard how do you come up with regulations that are clearly not flexible enough to move with the changing context yeah that's a good question but um the I think that the problem is that that the gray areas are tolerated maybe too much maybe it would be a good idea to more restrict these gray areas for instance when we take the Jews and the use of various and this study about the use in Berlin from from the GPA um when when it's it's public that they really use it in in a in a routine recognized the more I think written as they I wonder why why does not more debates or maybe um there should follow consequences strong visible consequences that that now it's it's uh less use and so that the law can catch up in way so we have to stand in the way frontal down the practice and it's speed the I the more questions the than In the meantime I just want you to say we had a discussion about them or it's all about what can we do against so veins in the web and I and I'm not too much into it but there are uh really innovative programmers trying to to develop a solution to the the to the 5th level of the confluence of Laplace's want to mention it and you can you can check it out and some projects owing on they they they tried to to to make new tools um which make it possible to have a social outlet without service or about raising functions so and it's always important to keep in mind that we have to regulate things because and you not everyone can become a techie but on the other hand we also can think of alternatives to all these uh big solutions so just maybe if someone wants to catch up the discussion from before because that's the way the same same problem and and that the is that you hear the word and things that for literary and you have to kind of rich