18 Feb
2021
18 Feb
'21
5:21 p.m.
Anhang verfügbar bis 20.03.2021
Dear members of the Ethics Committee,
maybe you heard that one of the SPARTA deliverables was rejected by the European
Commission because “the sources […] are not properly referenced”.
This is Deliverable 7.1 "AI systems threat analysis mechanisms and tools“.
I have attached the results of an automated similarity
(scribbr.com<http://scribbr.com>) that CEA has initated. You will see that a lot of
text is highlighted, not not necessarily indicating plagiarism: there are elements such as
the disclaimer, documents titles in the text and in the bibliography, quotations from the
GDP and other properly marked quotations.
At the beginning of the PDF there is a list to the domains where documents with similar
text were found, but it’s not possible to click on a link to see the full text of the
document.
Thibaud has already discussed the issue with the responsible editor from Tecnalia. They
have agree upon the following:
* some quotation marks (in Italics font) have been lost in the edition process ->
solution: we will use the quotation marks “” besides italics in all the quoted texts.
* in some cases the report reflects the taxonomies and techniques analysed too
literally -> solution: we will mark the quoted text and try to reduce the size of the
quotation as much as possible.
* some parts contain inputs from SPARTA partners that were too literal from their own
papers -> solution: we will mark the quoted text and provide the self-reference
So from my point of view there is no urgent need to intervene directly in this case
because a solution has been found.
However, it might be good if we write a reminder to all partners and remind them of the
following:
* that there are guidelines for good scientific practice that all SPARTA partners are
contractually obliged to follow.
* that there must be a good balance between what is already known and new knowledge in
each deliverable.
* that there must be a critical examination of third-party knowledge (out of many possible
threat frameworks, why did we adopt this one in SPARTA?).
* that the (financial) risk lies with the editor of a deliverable (in the case of D7.1 the
EC has withheld payment to the editor of the deliverable, but not to the contributing
partners).
This is perhaps a salutary shock that makes it clear that more careful quality control is
needed and that "minimum effort strategies" do not go undetected.
The EB has also started discussing whether deliverables should first go through the
automatic plagiarism check before they are sent to the internal reviewers.
If you agree with this approach, I would draft a message to all SPARTA parters which I
would submit to you in advance for your approval. If you feel it is necessary, we can also
schedule a meeting of the Ethics Committee to discuss how to proceed .
Best
Michael
Zum Laden
klicken<https://www.icloud.com/attachment/?u=https%3A%2F%2Fcvws.icloud-c…
1025719-Similarity-Check.pdf
23,1 MB
---
Dr. Michael Friedewald
Coordinator ICT Research
Competence Center Emerging Technologies
Fraunhofer-Institut für System- und Innovationsforschung ISI
Breslauer Straße 48 | 76139 Karlsruhe
fon: +49 721 6809-146 (-189, ass.) | fax: +49 721 6809-77-146
mailto:
michael.friedewald@isi.fraunhofer.de<mailto:michael.friedewald@isi.fraunhofer.de>
http://www.isi.fraunhofer.de
http://works.bepress.com/michael_friedewald/
Neue Veröffentlichungen:
Friedewald M., Schiffner S., Krenn S. (Eds.) (2021, im Erscheinen): Privacy and Identity
Management. 15th IFIP WG 9.2 9.6/11.7 11.6/SIG 9.2.2 International Summer School Maribor
Slovenia September 20-23 2020 Revised Selected Papers. Cham: Springer International (IFIP
Advances in Information and Communication Technology, 619).
Ammicht Quinn, R., Friedewald, M., Heesen, J., Krämer, N., Stapf, I. (Hrsg.) (2021, im
Erscheinen): Aufwachsen in überwachten Umgebungen: Interdisziplinäre Positionen zur
Privatheit und Datenschutz in Kindheit und Jugend. Baden-Baden: Nomos.
Martin, N.; Friedewald, M.; Schiering, I. et al. (2020): Die Datenschutz-Folgenabschätzung
nach Art. 35 DSGVO: Ein Handbuch für die Praxis. Stuttgart: Fraunhofer Verlag.
18 Feb
18 Feb
7:34 p.m.
Dear Michael,
I thank you for your email and, indeed, I was informed about the issue.
You are right about the need to be very strict about the quality of the
deliverables. I agree that deliverables go thru the automatic plagiarism
check before they are sent to the internal reviewers.
However, we have to be aware that some matters, as law for example, use
terms which are used more and more in legislation and doctrine as well
as in case laws. We also rewrite pieces of law to explain a position or
theory (and it's even more true when we do a state of the art document).
The anti-plagiarism will detect such quotes even if it's a usual way to
write reports and articles.
I wanted to highlight this particularity.
Best regards,
Jean-Marc
Jean-Marc Van Gyseghem
Research Director and scientific coordinator
Financial and Executive Manager of the advanced Master in ICT and Crids'
vocational trainings (@CRIDS, InfoSafe, DocSafe and DataSafe)
Research Centre Information, Law and Society (CRIDS)
T. +32 (0)81 72 52 08
F. +32 (0)81 72 52 02
jean-marc.vangyseghem(a)unamur.be <mailto:jean-marc.vangyseghem@unamur.be>
http://www.crids.eu <http://www.crids.eu>
CRIDS
Rempart de la Vierge 5 - 5000 Namur
Belgique
The Crids organises adult continuing formation dealing with new
technologies and communication called "JuriTIC" as well as vocational
trainings called "InfoSafe" "DocSafe".
Information and registration: http://www.crids.eu/formations
<http://www.crids.eu/formations>
This message is confidential. It may contain information that is covered
by legal professional privilege, work product immunity or other legal
rules. If you have received this communication in error, you must not
copy or distribute this message or any part of it or otherwise disclose
its contents to anyone. Please notify us by telephone or return E-mail,
and then delete this E-mail and any copies of it from your computer
system. Thank you.
Let's respect the environment together.
Only print this message if necessary!
Le 18/02/2021 à 18:21, Friedewald, Michael a écrit :
Anhang verfügbar bis 20.03.2021
Dear members of the Ethics Committee,
maybe you heard that one of the SPARTA deliverables was rejected by the European
Commission because “the sources […] are not properly referenced”.
This is Deliverable 7.1 "AI systems threat analysis mechanisms and tools“.
I have attached the results of an automated similarity
(scribbr.com<http://scribbr.com>) that CEA has initated. You will see that a lot of
text is highlighted, not not necessarily indicating plagiarism: there are elements such as
the disclaimer, documents titles in the text and in the bibliography, quotations from the
GDP and other properly marked quotations.
At the beginning of the PDF there is a list to the domains where documents with similar
text were found, but it’s not possible to click on a link to see the full text of the
document.
Thibaud has already discussed the issue with the responsible editor from Tecnalia. They
have agree upon the following:
* some quotation marks (in Italics font) have been lost in the edition process ->
solution: we will use the quotation marks “” besides italics in all the quoted texts.
* in some cases the report reflects the taxonomies and techniques analysed too
literally -> solution: we will mark the quoted text and try to reduce the size of the
quotation as much as possible.
* some parts contain inputs from SPARTA partners that were too literal from their
own papers -> solution: we will mark the quoted text and provide the self-reference
So from my point of view there is no urgent need to intervene directly in this case
because a solution has been found.
However, it might be good if we write a reminder to all partners and remind them of the
following:
* that there are guidelines for good scientific practice that all SPARTA partners are
contractually obliged to follow.
* that there must be a good balance between what is already known and new knowledge in
each deliverable.
* that there must be a critical examination of third-party knowledge (out of many
possible threat frameworks, why did we adopt this one in SPARTA?).
* that the (financial) risk lies with the editor of a deliverable (in the case of D7.1
the EC has withheld payment to the editor of the deliverable, but not to the contributing
partners).
This is perhaps a salutary shock that makes it clear that more careful quality control is
needed and that "minimum effort strategies" do not go undetected.
The EB has also started discussing whether deliverables should first go through the
automatic plagiarism check before they are sent to the internal reviewers.
If you agree with this approach, I would draft a message to all SPARTA parters which I
would submit to you in advance for your approval. If you feel it is necessary, we can also
schedule a meeting of the Ethics Committee to discuss how to proceed .
Best
Michael
Zum Laden
klicken<https://www.icloud.com/attachment/?u=https%3A%2F%2Fcvws.icloud-c…
1025719-Similarity-Check.pdf
23,1 MB
---
Dr. Michael Friedewald
Coordinator ICT Research
Competence Center Emerging Technologies
Fraunhofer-Institut für System- und Innovationsforschung ISI
Breslauer Straße 48 | 76139 Karlsruhe
fon: +49 721 6809-146 (-189, ass.) | fax: +49 721 6809-77-146
mailto:
michael.friedewald@isi.fraunhofer.de<mailto:michael.friedewald@isi.fraunhofer.de>
http://www.isi.fraunhofer.de
http://works.bepress.com/michael_friedewald/
Neue Veröffentlichungen:
Friedewald M., Schiffner S., Krenn S. (Eds.) (2021, im Erscheinen): Privacy and Identity
Management. 15th IFIP WG 9.2 9.6/11.7 11.6/SIG 9.2.2 International Summer School Maribor
Slovenia September 20-23 2020 Revised Selected Papers. Cham: Springer International (IFIP
Advances in Information and Communication Technology, 619).
Ammicht Quinn, R., Friedewald, M., Heesen, J., Krämer, N., Stapf, I. (Hrsg.) (2021, im
Erscheinen): Aufwachsen in überwachten Umgebungen: Interdisziplinäre Positionen zur
Privatheit und Datenschutz in Kindheit und Jugend. Baden-Baden: Nomos.
Martin, N.; Friedewald, M.; Schiering, I. et al. (2020): Die
Datenschutz-Folgenabschätzung nach Art. 35 DSGVO: Ein Handbuch für die Praxis. Stuttgart:
Fraunhofer Verlag.
19 Feb
19 Feb
7:09 a.m.
Dear Jean-Marc,
thanks for your message. Yes we can see there is a lot of highlighted text in Chapter 4 as
well but the patterns are different. There are quite a lot of well referenced quotes in
that chapter and I also fully see your point that in law there are standing definitions
that do not improve when they are „creatively“ reformulated. Apart from that I am not
completely clear what the tool does. In your chapter it highlights quite a lot of very
short text segments (see for instance the bullet point list on page 83 with number 32 in
the circle: for me it is hard to imagine what kind of copy and paste should have happened
there…) In short: I do not see an issue in that chapter.
Your points regarding quality control have been taken.
Best
Michael
Am 18.02.2021 um 20:34 schrieb Jean-Marc VAN GYSEGHEM
<jean-marc.vangyseghem@unamur.be<mailto:jean-marc.vangyseghem@unamur.be>>:
Dear Michael,
I thank you for your email and, indeed, I was informed about the issue.
You are right about the need to be very strict about the quality of the deliverables. I
agree that deliverables go thru the automatic plagiarism check before they are sent to the
internal reviewers.
However, we have to be aware that some matters, as law for example, use terms which are
used more and more in legislation and doctrine as well as in case laws. We also rewrite
pieces of law to explain a position or theory (and it's even more true when we do a
state of the art document). The anti-plagiarism will detect such quotes even if it's a
usual way to write reports and articles.
I wanted to highlight this particularity.
Best regards,
Jean-Marc
Jean-Marc Van Gyseghem
Research Director and scientific coordinator
Financial and Executive Manager of the advanced Master in ICT and Crids' vocational
trainings (@CRIDS, InfoSafe, DocSafe and DataSafe)
Research Centre Information, Law and Society (CRIDS) <gbccbojhpcnbafla.png>
T. +32 (0)81 72 52 08
F. +32 (0)81 72 52 02
jean-marc.vangyseghem@unamur.be<mailto:jean-marc.vangyseghem@unamur.be>
http://www.crids.eu<http://www.crids.eu/>
CRIDS
Rempart de la Vierge 5 - 5000 Namur
Belgique
The Crids organises adult continuing formation dealing with new technologies and
communication called "JuriTIC" as well as vocational trainings called
"InfoSafe" "DocSafe".
Information and registration: http://www.crids.eu/formations
This message is confidential. It may contain information that is covered by legal
professional privilege, work product immunity or other legal rules. If you have received
this communication in error, you must not copy or distribute this message or any part of
it or otherwise disclose its contents to anyone. Please notify us by telephone or return
E-mail, and then delete this E-mail and any copies of it from your computer system. Thank
you.
Let's respect the environment together.
Only print this message if necessary!
Le 18/02/2021 à 18:21, Friedewald, Michael a écrit :
Anhang verfügbar bis 20.03.2021
Dear members of the Ethics Committee,
maybe you heard that one of the SPARTA deliverables was rejected by the European
Commission because “the sources […] are not properly referenced”.
This is Deliverable 7.1 "AI systems threat analysis mechanisms and tools“.
I have attached the results of an automated similarity
(scribbr.com<http://scribbr.com><http://scribbr.com><http://s…)
that CEA has initated. You will see that a lot of text is highlighted, not not necessarily
indicating plagiarism: there are elements such as the disclaimer, documents titles in the
text and in the bibliography, quotations from the GDP and other properly marked
quotations.
At the beginning of the PDF there is a list to the domains where documents with similar
text were found, but it’s not possible to click on a link to see the full text of the
document.
Thibaud has already discussed the issue with the responsible editor from Tecnalia. They
have agree upon the following:
* some quotation marks (in Italics font) have been lost in the edition process ->
solution: we will use the quotation marks “” besides italics in all the quoted texts.
* in some cases the report reflects the taxonomies and techniques analysed too
literally -> solution: we will mark the quoted text and try to reduce the size of the
quotation as much as possible.
* some parts contain inputs from SPARTA partners that were too literal from their own
papers -> solution: we will mark the quoted text and provide the self-reference
So from my point of view there is no urgent need to intervene directly in this case
because a solution has been found.
However, it might be good if we write a reminder to all partners and remind them of the
following:
* that there are guidelines for good scientific practice that all SPARTA partners are
contractually obliged to follow.
* that there must be a good balance between what is already known and new knowledge in
each deliverable.
* that there must be a critical examination of third-party knowledge (out of many possible
threat frameworks, why did we adopt this one in SPARTA?).
* that the (financial) risk lies with the editor of a deliverable (in the case of D7.1 the
EC has withheld payment to the editor of the deliverable, but not to the contributing
partners).
This is perhaps a salutary shock that makes it clear that more careful quality control is
needed and that "minimum effort strategies" do not go undetected.
The EB has also started discussing whether deliverables should first go through the
automatic plagiarism check before they are sent to the internal reviewers.
If you agree with this approach, I would draft a message to all SPARTA parters which I
would submit to you in advance for your approval. If you feel it is necessary, we can also
schedule a meeting of the Ethics Committee to discuss how to proceed .
Best
Michael
Zum Laden
klicken<https://www.icloud.com/attachment/?u=https%3A%2F%2Fcvws.icloud-c…
1025719-Similarity-Check.pdf
23,1 MB
---
Dr. Michael Friedewald
Coordinator ICT Research
Competence Center Emerging Technologies
Fraunhofer-Institut für System- und Innovationsforschung ISI
Breslauer Straße 48 | 76139 Karlsruhe
fon: +49 721 6809-146 (-189, ass.) | fax: +49 721 6809-77-146
mailto:
michael.friedewald@isi.fraunhofer.de<mailto:michael.friedewald@isi.fraunhofer.de><mailto:michael.friedewald@isi.fraunhofer.de><mailto:michael.friedewald@isi.fraunhofer.de>
http://www.isi.fraunhofer.de<http://www.isi.fraunhofer.de/>
http://works.bepress.com/michael_friedewald/
Neue Veröffentlichungen:
Friedewald M., Schiffner S., Krenn S. (Eds.) (2021, im Erscheinen): Privacy and Identity
Management. 15th IFIP WG 9.2 9.6/11.7 11.6/SIG 9.2.2 International Summer School Maribor
Slovenia September 20-23 2020 Revised Selected Papers. Cham: Springer International (IFIP
Advances in Information and Communication Technology, 619).
Ammicht Quinn, R., Friedewald, M., Heesen, J., Krämer, N., Stapf, I. (Hrsg.) (2021, im
Erscheinen): Aufwachsen in überwachten Umgebungen: Interdisziplinäre Positionen zur
Privatheit und Datenschutz in Kindheit und Jugend. Baden-Baden: Nomos.
Martin, N.; Friedewald, M.; Schiering, I. et al. (2020): Die Datenschutz-Folgenabschätzung
nach Art. 35 DSGVO: Ein Handbuch für die Praxis. Stuttgart: Fraunhofer Verlag.
--
bodies.ethics-committee mailing list
bodies.ethics-committee@server.sparta.eu<mailto:bodies.ethics-committee@server.sparta.eu>
http://server.sparta.eu/cgi-bin/mailman/listinfo/bodies.ethics-committee
1694
days inactive
1695
days old
bodies.ethics-committee@internal.sparta.eu
2 comments
2 participants
participants (2)
-
Friedewald, Michael -
Jean-Marc VAN GYSEGHEM