Hey all, below are the questions we came up with after our Tuesday Meeting and which was asked to the project director here at York. In the next days we'll send around an overview of the answers, since the conversation was informal, ie off the record.

Introduction

INDECT is a document analysis tool through Natural Language Processing to create identities of people, their relations and events at which they participate based on a data source. The following questions will be asked to Suresh Manandhar of CS UoY, the possibility of this interview is given only if answers are strictly off the record. The questions have been formulated by members of the Free Culture society at the University of York.


Questions


1) Doesn’t the Indect Project practically change the architecture of

the internet in a restrictive way, if individuals habits and behaviour

can be tracked and correlated simultaneously across the entire

network?

This project makes further development in making code the law.

This can be seen for example in an “event paired with a salient entity”

19/41 4.1

“co-reference resolution” 11/41 4.1

“Tag prediction for conversion” Supervised machine learning, translating

the internet 27/41 4.1


2)How will you avoid and elimate from a database false positives,

spamming and spurious responses and dubious accuracy of statements

e.g. seperating jokes/parody from serious threats, “Google bombing” or

malcious campaigns to ruin an individuals reputation?

For example “Joking and Spartaco on Twitter” 33/41 4.1

“Person inherits properties from agent and object” -hasRelative 4.1 31/41


3) Collect(monitor) information on everybody even innocent?

Achieved by ACE + Ontology = Extensibility

ACE characterized by:

-# of different entity types

-relations between them

-events in which entities participate

“Bidirectional many-to-many type” 33/41 4.1


4) Even though INDECT may be guided by rigorous ethics, how can you

ensure that others do not replicate a less ethical system with the

information and algorithms you will publish during the project? In

that way the technologies developed could be contributing to a

supression of legitimate protest, e.g. Chinese police forces using the

project to improve the effectiveness of monitoring human rights activists.


5) For how long will data of people who are not suspects or eliminated

will be kept for? The Data Protection Act mandates that data must be

only kept for specific purposes and deleted when no longer

required. But if data is deleted regularly to meet this legal

requirement, how much use will the system be in making correlations

with past behaviour?

“Fact of highlighting no legal consequences, no permanent record would

be kept” 2/2 Ethical focused


6)Why does the Ethical issues paper has 8 instances of “Police” and not even

one explaining the system from the users point of view?


7)Would copyright infringement be a “justified reasons for

interference”? 8/30 D0.5

“Registration and storage of only situations related to a threat,

collecting ALL(clarify) monitored situations” 8/30 D0.5


8)Isn’t transparency, correctness and privacy partly obscured or brought

into question when you say:

“No issues that could impact negatively upon

-Law enforcement capability

-National Security

-Public Safety

-Organisational reputation

should be published in the public domain” 10/30 D0.5

leading to:


9) Why is this project then being funded as R&D and not as a security

project? In the intelligence community, are there not already

monitoring systems for serious crime which partially accomplish

INDECT’s end goals or more? e.g. Echelon. It may be seen as a waste of

scarce

research money to redevelop this technology from the ground up when

the intelligence services have already built up significant experience

in combating serious crime e.g. terrorism through electronic surveillance.


10)Doesn’t automatic detection and recognition of criminal behaviour,

real time alert have a possibility of targeting someone before a

crime is committed on the basis of statistical likelihoods? 11/30 D0.5


11) Do “hackers” include people who incur in copyright infringement?

12/30 D0.5


12)Why would Facebook and other social networks make data

available voluntarily, given the negative publicity they have already

suffered

over privacy problems, and the potential for further backlash from

users? Is there a possibility that INDECT involvement may merely be

hidden away in a privacy policy with a auto check-box which no-one

reads, and thus information might be collected without user’s real

informed consent on such social networks? 12/30 D0.5


13)Does “criminal actions on the internet” include filesharing of

copyrighted content, distribution of legitimate security engineering

tools etc? 12/30 D0.5


14) Isn’t this project aimed at cataloguing a majority of people who

have nothing to hide? Wouldn’t all serious criminal gangs and

terrorist use encryption techniques, which you say will not be

scrutinized for data? Alternatively, once real systems based on INDECT

are deployed and publicised, is it not more likely that serious

criminals using plain-text would shift to custom codes e.g. using

special slang or Internet “lol-speak” that would confuse natural language

processing algorithms?


15)Just because I might be a suspect for a possible crime does that

mean the police should profile all my life, without specific evidence

linking me to the crime?

“Profiling suspected figureheads even with anonymous identities […] is

expected” 12/30 D0.5


16) Why is INDECT being funded as a research project but treated as a

security one. Will it comply with the Data Protection Act, so that

someone considered a suspect and then found innocent can check the

data that is being held about him, or will it be classified under

prevention and detection of crime?


17) ”INDECT researchers do not hold and do not intend to hold data

normally requiring a warrant or court order”(13/30 D0.5), what about

in the real system? Is a checks and balance system, e.g. judicial

review being designed as a core part of the project to monitor its

final usage?


18) “No personal data is being processed […] at current phase of

project realisation” 17/30 D0.5 isn’t conflicting with what is being

written regarding video monitoring systems installed in the Warsaw

Metro(28/30 D0.5)?


19)Could police partners be the Chinese police? 18/30 D0.5

So in the end the police will glue this project to reality?


20)Considering the ease of accessing information, who would audit logs

of INDECT lookups to make sure the threat was proportionate? For

example, the RIPA survillance powers in the UK were initially

introduced to combat similar threats to INDECT, e.g. terrorism and

paedophilia, but ended up being used to allow local councils to

monitor minor offences e.g. litter/dog fouling? Is there not a risk

that with target-driven policing and the mass monitoring capacity like

INDECT provide, it will ultimately be used to “trawl” for suspects as

a first stage, rather than a tool for targeted investigation? FAQ


21) “Detection of websites, blogs, forums that promote illegal

activities” include copyrighted content? 9/55 4.3


22) Does “email analysis” as specified in 34/55 4.3 mean that

“classified data”(6/7 FAQ) information will be analysed?


23) Can the analysis of “sequential patterns of network events” (42/55

4.3) be used for tracking for example P2P users sharing content? Is

using P2P networks even for legal purposes e.g. distributing Linux DVD

images, and use of anonymiser networks e.g. Tor going to eventually

become a risk factor and potential indicator of wrong-doing?


24)Doesn’t the phrase “INDECT is a research project, not an

implementation project […] rather than production phase product”

(18/30 4.3) conflict with “One of the main criteria for accepting

financial support from the EC for research done within INDECT and

similar project are prospects for practical implementation of research

outcomes” (3/7 FAQ)?


25)Doesnt avoiding interviews on the record spark doubts about the

intentions of the project itself? If monitoring networks based upon

INDECT demand transparency from citizens, isn’t it fair the public

receive full transparency from those involved in developing and

implementing these systems?




_______________________________________________
Discuss mailing list
[email protected]
http://freeculture.org/cgi-bin/mailman/listinfo/discuss
FAQ: http://wiki.freeculture.org/Fc-discuss

Reply via email to