
PHD Student @ UniPaderborn | Responsible AI for Biometrics | MSc Cse @UniFreiburg | Prev @MPI_IS | @FraunhoferISE | #MPISP
Germany
Joined September 2012
- Tweets45
- Following444
- Followers73
- Likes490
24 hrs. 2 people. 1 mission.
At the {Tech: Europe} Karlsruhe AI Hackathon, we built 6sense — a real-time multimodal AI assistant for the visually impaired.
Powered by @MistralAI , @elevenlabs , and @runpod .
Proud to build this with @deepukreddy .
Runner-up 🥈
#AIforGood
🎓✨ A small personal update from my side:
I officially finished my Master's degree @UniFreiburg!
I am very excited to join @MPI_IS as a research intern under the supervision of Prof. Dr. Antonio Orvieto @orvieto_antonio! I will be working on long-range music generation. 🎶🚀
1
6
452
During my master's, I am very grateful to have worked under great supervisors: @asia_biega, @_LinKyi, @MariaAleBravo, Mohammadreza Zolfaghari, and @ThomasBrox. Thank you for playing a huge role in my life! 🙏✨
1
3
201
Sushil Ammanaghatta Shivakumar retweeted
🏆 AWARDED PAPER ALERT: deceptive design in the context of GDPR's legitimate interest. More info & important quotes below:
The paper "Investigating Deceptive Design in GDPR's Legitimate Interest," authored by @_LinKyi, @asia_biega, @franziroesner, @Cristianapt, Sushil Ammanaghatta Shivakumar, and Frederike Zufall, was published in 2023 and has recently won the Stefano Rodotà Data Protection Award 2024.
Deceptive design & dark patterns are some of the topics of my own Ph.D. research, and this paper brings a very important empirical contribution to the field by analyzing deceptive design practices in the context of GDPR's legitimate interest.
Here are some important quotes:
"The use of legitimate interest is exploited through deceptive design practices both at UI and linguistic levels. From our findings, we noticed that the implementation of this legal ground has many UI and linguistic deceptive design elements so far unreported in the literature of deceptive designs. In particular, more attention should be paid to deceptive designs that go beyond the UI. Complex deceptive designs often combine multiple dubious practices in one interface, as our study found with linguistic and UI deceptive designs being used when presenting legitimate interest. (page 12)
-
"Neither participants nor the law approve of advertising purposes under the legitimate interests legal basis, despite its common usage. As observed in Section 2.1, the use of the opentextured term, “legitimate interests” allows controllers to base the processing of personal data on a balancing decision when indeed their own legitimate interests outweigh those of the data subjects. This is necessary because the law (Art. 6(1) GDPR) may not foresee all potential scenarios that may occur in practice." (page 13)
-
"Lack of enforcement. The function of “legitimate interests” as a legal basis for cases that were not foreseen by the law has one important consequence: the decision on whether legitimate interests actually outweigh the protection of personal data, and thus are a valid legal basis, is reliant on the case-by-case assessment of executive decisions by DPAs and judicial findings of courts. This means that this huge margin of interpretation ultimately depends on assessing these cases in practice. Otherwise, as our studies show, it may be exploited to the detriment of data subjects’ rights." (page 14)
-
As many privacy professionals have already realized, there is an overreliance on legitimate interest to confer lawfulness to data processing. This paper sheds light on the fact that this overreliance might be, in fact, abusive, as it often deploys dark patterns.
*Prof. @Cristianapt - one of the paper's co-authors - will be my guest on the podcast next week, together with Prof. @hartzog. We will discuss deceptive design and more. Register (link in the comments) and join us live.
Read the paper below (link⬇️)
Sushil Ammanaghatta Shivakumar retweeted
🎓#PhD👏 Félicitations à Swarn PRIYA👩🎓doctorante @Univ_CotedAzur au sein de l'équipe Stamp @inria_sophia pour son travail de thèse et sa soutenance !
📚 #Sécurité #Logiciel
En savoir ➕ sur elle ➡️ bit.ly/3v5prs4
➡️ team.inria.fr/stamp/fr/
Sushil Ammanaghatta Shivakumar retweeted
We're looking for a research assistant (HiWi) to support us in dataset & software development in the @HorizonFINDHR project. Required: strong Python programming skills+English fluency+enrolled at a German uni. Remote possible. Details & how to apply:
mpg.de/21115213/student-or-r…
🙏 RT
Best ad of Duolingo??
youtube.com/watch?v=P6FORpg0…
1
55
Sushil Ammanaghatta Shivakumar retweeted
Today, @BeatriceLansbe1 talks to @asia_biega from Max Planck Institute for Security and Privacy, about #data minimisation, data protection laws, #ethical issues in the data collection, and responsible computing.
#computerscience #privacy
pod.fo/e/1780ff
Sushil Ammanaghatta Shivakumar retweeted
"Investigating Deceptive Design in GDPR’s Legitimate Interest" with @_LinKyi @shilsuba @Cristianapt @franziroesner and F. Zufall
dl.acm.org/doi/10.1145/35445…
#CHI2023
3/3
Sushil Ammanaghatta Shivakumar retweeted
Looking at all the farewell pictures which juniors have been posting, I wish we could also have a farewell which we never had. #batchof2021 – at Brooklyn, NY




















