[Lilab] Fwd: [Science_faculty_all] [Action Required] Report the storage or usage of a corrupted research dataset (LAION-5B)

[Wei Vivian Li]

FYI.


---------- Forwarded message ---------
From: CNAS Dean <CNAS_DEAN at ucr.edu>
Date: Wed, Jan 10, 2024 at 2:05 PM
Subject: [Science_faculty_all] [Action Required] Report the storage or
usage of a corrupted research dataset (LAION-5B)
To: science_faculty_all at lists.ucr.edu <science_faculty_all at lists.ucr.edu>,
cnas_staff at lists.ucr.edu <cnas_staff at lists.ucr.edu>
Cc: PHYSCIdean <physcidean at ucr.edu>, Jennifer V Farias <
jennifer.farias at ucr.edu>, Joann D Young <joann.young at ucr.edu>, Agdean <
agdean at ucr.edu>, LifeSciDean <lifescidean at ucr.edu>, CNAS Dean <
CNAS_DEAN at ucr.edu>, Studentdeancnas <studentdeancnas at ucr.edu>


CNAS Colleagues,



I am forwarding this URGENT message to all. Please read and follow ‘action’
steps below, if warranted.





Peter W. Atkinson, Ph.D.
Interim Dean, College of Natural & Agricultural Sciences

Professor of Genetics

Department of Entomology

University of California, Riverside
Office: (951) 827-3101  |  Email: CNAS_Dean at ucr.edu

Executive Assistant: Rachel Alvarez  |  Office: (951) 827-4597  |
rachel.alvarez at ucr.edu
[image: A blue and yellow text on a black background Description
automatically generated]








*Subject: *[Action Required] Report the storage or usage of a corrupted
research dataset (LAION-5B)

*Sent on behalf of CISO Dewight Kramer*



UCR Vice Chancellors and Deans,



I am writing to request your assistance to comply with a University of
California Office of the President (UCOP) mandate to determine whether a
corrupted research dataset, referred to as LAION-5B, is being stored or
utilized by anyone at UCR. *Please share this urgent message with faculty
and anyone else who may be using AI data.*



*Issue and Urgency*

A publicly available, open-source dataset widely used to train artificial
intelligence (AI) systems, referred to as LAION-5B, is believed to contain
links to thousands of images of child sexual abuse material, including
child pornography, as defined under 18 U.S.C. § 2256(8). This dataset comes
from the Large-scale Artificial Intelligence Open Network, or LAION, a
non-profit organization that creates open-source tools for machine learning.



UC faculty and researchers may have downloaded the images of this dataset
for legitimate research purposes (and without knowledge that it contained
child sexual abuse material), potentially storing child sexual abuse
material on UC information technology (IT) resources. US and international
laws prohibit the possession and distribution of such material. Therefore,
it is urgent that anyone who has downloaded or utilized the LAION-5B
dataset immediately contact UCR’s Chief Information Security Officer (CISO)
at security at ucr.edu.



Action Required

Any member of the UCR community – faculty, staff, postdoctoral fellows, or
graduate or undergraduate students – who may have downloaded the LAION-5B
dataset must take the following steps:

1.     *Immediately notify the UCR CISO at **security at ucr.edu*
<security at ucr.edu> to make the Information Security Office aware of the
suspected storage or utilization of this dataset.

2.     *Do not* *attempt to delete or modify the dataset*.

3.     *Refrain from any further use of the dataset*.



*It is important to note that individuals who downloaded or have access to
the dataset are not at fault*. This dataset was corrupted, and these steps
are being required by UCOP to protect the UCR community. UCR CISO and legal
counsel may need to meet with affected individuals to gather the
information needed to take appropriate action. We thank you in advance for
your cooperation.



Additional Information

On December 21, 2023, UCOP requested that all UC campuses determine whether
the corrupted research dataset, LAION-5B, is being stored or utilized by
anyone on campus. UCR has been asked to report its findings to UCOP no
later than January 19, 2024.



Upon receiving this request, the UCR Information Security Office
immediately reached out to key staff who provide research computing support
to UCR researchers. At this time, there are no clear indicators that anyone
on campus has downloaded the dataset in question; however, your support is
required to ensure that this is the case.



Thank you for your understanding and cooperation as UCR continues to
address this serious issue.



Sincerely,



*Dewight F. Kramer*

*Chief Information Security Officer*

Information Security Office

Information Technology Solutions

University of California, Riverside


_______________________________________________
Science_faculty_all mailing list
Science_faculty_all at lists.ucr.edu
https://lists.ucr.edu/mailman/listinfo/science_faculty_all
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ucr.edu/pipermail/lilab/attachments/20240110/40830619/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 9419 bytes
Desc: not available
URL: <https://lists.ucr.edu/pipermail/lilab/attachments/20240110/40830619/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 9939 bytes
Desc: not available
URL: <https://lists.ucr.edu/pipermail/lilab/attachments/20240110/40830619/attachment-0003.png>