pith. sign in

arxiv: 1907.09334 · v1 · pith:6A6MXKAAnew · submitted 2019-07-17 · 💻 cs.HC

LinTO : Assistant vocal open-source respectueux des donn\'ees personnelles pour les r\'eunions d'entreprise

Jean-Pierre Lorr\'e , Isabelle Ferran\'e (IRIT) , Francisco Madrigal (CIMAT) , Michalis Vazirgiannis (LIX) , Christophe Bourguignat This is my paper

Pith reviewed 2026-05-24 20:32 UTC · model grok-4.3

classification 💻 cs.HC
keywords open-source assistantGDPR complianceenterprise meetingsconversational AIvoice assistantdata privacymeeting facilitationminute writing
0
0 comments X

The pith

LinTO is presented as the first open-source enterprise assistant designed to comply with GDPR requirements.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces LinTO, a conversational assistant intended to support company employees during meetings. It consists of an interactive device equipped with microphones, a screen, and a 360 degree camera that can control the room, query the company's information system, facilitate meetings, and aid minute writing. The system follows an open model that respects private data. A sympathetic reader would care because it claims to solve the tension between useful AI assistance and strict data protection rules in business environments.

Core claim

LinTO is the first open-source enterprise's assistant designed to comply with the GDPR requirements, using an interactive device with microphones, a screen and a 360 camera to provide meeting support while respecting private data through an open distribution model.

What carries the argument

The LinTO interactive device with microphones, screen and 360 camera that enables voice control, room management, information queries, meeting facilitation and minute-writing assistance under an open GDPR-compliant model.

If this is right

  • The device can control the room and query the company's information system during meetings.
  • It helps facilitate meetings and provides an environment to aid minute writing.
  • Data is protected through an open-source distribution model that meets GDPR rules.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the hardware works as described, similar open devices could be adapted for other privacy-sensitive enterprise tasks.
  • Verification of actual GDPR compliance would require external audits beyond the project description.
  • Integration with existing company systems might depend on how the open model handles proprietary data interfaces.

Load-bearing premise

The described hardware and open model actually deliver both functional meeting assistance and full GDPR compliance.

What would settle it

A real-world deployment in company meetings that either fails to provide the claimed assistance functions or leaks personal data in violation of GDPR.

Figures

Figures reproduced from arXiv: 1907.09334 by Christophe Bourguignat, Francisco Madrigal (CIMAT), Isabelle Ferran\'e (IRIT), Jean-Pierre Lorr\'e, Michalis Vazirgiannis (LIX).

Figure 1
Figure 1. Figure 1: Chaîne de traitement en mode commande. 4.1 Détection du « hotword » Le mode commande est déclenché à l’aide d’un mécanisme de réveil qui consiste à reconnaître un ou plusieurs mots spécifiques prononcés par l’utilisateur afin de s’activer. Dans cette approche la détection d’un mot-clef (hotword) prononcé par l’utilisateur précède une commande ou une question. Le dispositif sait que la phrase suivant le hot… view at source ↗
read the original abstract

This paper presents the first results of the PIA "Grands D\'efis du Num\'erique" research project LinTO. The goal of this project is to develop a conversational assistant to help the company's employees, particularly during meetings. LinTO is an interactive device equipped with microphones, a screen and a 360$^\circ$ camera, which allows to control the room, query company's information system, helps facilitate the meeting and provides an environment to aid minute writing. Distributed according to an open model that respects private data LinTO is the first open-source enterprise's assistant designed to comply with the GDPR requirements.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The manuscript presents the LinTO project as the first results of the PIA 'Grands Défis du Numérique' initiative. It describes an interactive vocal assistant device equipped with microphones, a screen, and a 360° camera for use in enterprise meetings. The system is intended to control the room, query the company's information system, facilitate meetings, and support minute writing. The paper claims that LinTO follows an open model that respects private data and is the first open-source enterprise assistant designed to comply with GDPR requirements.

Significance. If the claims of functionality and GDPR compliance hold with supporting technical evidence, the work could contribute to privacy-preserving conversational agents in professional settings by providing an open-source alternative that addresses European data protection regulations. However, the current high-level description without implementation details, evaluations, or compliance verification limits its assessed impact to that of a project announcement rather than a substantiated technical contribution.

major comments (1)
  1. [Abstract] Abstract: The central claim that 'LinTO is the first open-source enterprise's assistant designed to comply with the GDPR requirements' is not supported by any architecture description, data-flow details, encryption or consent mechanisms, or explicit audit against GDPR articles (such as Art. 5, 25, or 32). The manuscript supplies only a high-level list of hardware components and intended features, leaving the compliance assertion as an unverified design goal rather than a demonstrated property.
minor comments (1)
  1. The title is in French while the abstract is written in English; consider adding a consistent bilingual abstract or clarifying the intended language for the full manuscript.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and for highlighting the need to substantiate claims in the abstract. We address the single major comment below and indicate the revision we will make.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that 'LinTO is the first open-source enterprise's assistant designed to comply with the GDPR requirements' is not supported by any architecture description, data-flow details, encryption or consent mechanisms, or explicit audit against GDPR articles (such as Art. 5, 25, or 32). The manuscript supplies only a high-level list of hardware components and intended features, leaving the compliance assertion as an unverified design goal rather than a demonstrated property.

    Authors: We agree that the present manuscript is a high-level project overview and does not contain the requested architecture diagrams, data-flow specifications, encryption details, consent mechanisms, or article-by-article GDPR audit. The claim in the abstract reflects the project's stated design objective (open-source distribution with explicit attention to private data) rather than a completed technical verification. Because this paper is positioned as the first public description of the PIA-funded effort, we will revise the abstract to read that LinTO 'is designed to comply with GDPR requirements' and will remove the stronger phrasing that could be read as an already-demonstrated property. If space permits, we will also add a sentence directing readers to forthcoming technical deliverables for the detailed compliance analysis. revision: yes

Circularity Check

0 steps flagged

No circularity; descriptive project paper with no derivations

full rationale

The manuscript is a high-level project announcement describing hardware, an open model, and a design goal of GDPR compliance. It contains no equations, no fitted parameters, no derivation chain, and no self-citations that bear load on any result. The 'first' claim is an unsupported assertion, not a reduction of one quantity to another by construction. All enumerated circularity patterns are absent; the paper is self-contained as a non-mathematical description.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities in the scientific sense; the paper describes a practical system development project without mathematical or theoretical components.

pith-pipeline@v0.9.0 · 5665 in / 938 out tokens · 19296 ms · 2026-05-24T20:32:52.489210+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

13 extracted references · 13 canonical work pages · 1 internal anchor

  1. [1]

    LinTO : Assistant vocal open-source respectueux des données personnelles pour les réunions d’entreprise Jean-Pierre Lorré1, Isabelle Ferrané2, Francisco Madrigal3, Michalis Vazirgiannis4, Christophe Bourguignat5 1LINAGORA 2IRIT 3LAAS-CNRS 4Laboratoire d’Informatique de l’Ecole Polytechnique 5Zelros jplorre@linagora.com, isabelle.ferrane@irit.fr, jfmadrig@...

    work page 2022

  2. [2]

    en croissance annuelle de 32% entre 2015 et

    work page 2015

  3. [3]

    et entraîné sur le corpus AMI [6]. Les évaluations montrent que l’encodeur proposé permet d’obtenir des groupes d’utterances proches des solutions humaines, et généralement meilleurs que ceux retournés par d’autres systèmes de l'état de l’art [14]. Par ailleurs, le corpus AMI est en cours de traduction via une approche de crowdsourcing afin de pouvoir dév...

    work page 2018

  4. [4]

    Bahdanau et al, Neural machine translation by jointly learning to align and translate, arXiv preprint arXiv:1409.0473 (2014)

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  5. [5]

    Chen L. et al. (2006) VACE Multimodal Meeting Corpus. In: Renals S., Bengio S. (eds) Machine Learning for Multimodal Interaction. MLMI

    work page 2006

  6. [6]

    Bengio, Reaching new records in speech recognition, Mars 2017, https://www.ibm.com/blogs/watson/2017/03/reaching-new-records-in-speech-recognition/

    J. Bengio, Reaching new records in speech recognition, Mars 2017, https://www.ibm.com/blogs/watson/2017/03/reaching-new-records-in-speech-recognition/

    work page 2017

  7. [7]

    Galliano et al, The ester 2 evaluation campaign for the rich transcription of French radio broadcasts, Interspeech-2009, 2583-2586

    work page 2009

  8. [8]

    Global Intelligent Virtual Assistant Market, juin 2018 https://www.transparencymarketresearch.com/pressrelease/intelligent-virtual-assistant-industry.htm

    work page 2018

  9. [9]

    Povey et al, The Kaldi Speech Recognition Toolkit, Workshop on Automatic Speech Recognition and Understanding, 2011

    D. Povey et al, The Kaldi Speech Recognition Toolkit, Workshop on Automatic Speech Recognition and Understanding, 2011

    work page 2011

  10. [10]

    Ratner et al, Rapid Training Data Creation with Weak Supervision, Proceedings of the VLDB Endowment, 11(3), 269-282, 2017

    A. Ratner et al, Rapid Training Data Creation with Weak Supervision, Proceedings of the VLDB Endowment, 11(3), 269-282, 2017

    work page 2017

  11. [11]

    Renals et al, ROCKIT: roadmap for conversational interaction technologies

    S. Renals et al, ROCKIT: roadmap for conversational interaction technologies. Proceedings of the 2014 Workshop on Roadmapping the Future of Multimodal Interaction Research including Business Opportunities and Challenges. ACM,

    work page 2014

  12. [12]

    Sercan O et al

    A. Sercan O et al. Convolutional Recurrent Neural Networks for Small-Footprint Keyword Spotting, Interspeech 2017

    work page 2017

  13. [13]

    Shang et al, Energy-based Self-attentive Learning of Abstractive Communities for Spoken Language Understanding, arXiv:1904.09491 [cs.CL]

    G. Shang et al, Energy-based Self-attentive Learning of Abstractive Communities for Spoken Language Understanding, arXiv:1904.09491 [cs.CL]

    work page arXiv 1904