Skip to main navigation Skip to main content
info(at)avldigital.remove-this.de
Deutsch English
Narrow Search
Search narrowed by
Last searches

Results for *

Displaying results 1 to 3 of 3.

  1. Strategic behavior and AI training data
    Author: Peukert, Christian; Abeillon, Florian; Haese, Jérémie; Kaiser, Franziska; Staub, Alexander
    Published: April 2024
    Publisher:  CESifo, Munich, Germany

    Human-created works represent critical data inputs to artificial intelligence (AI). Strategic behaviour can play a major role for AI training datasets, be it in limiting access to existing works or in deciding which types of new works to create or... more

    Access:
    Verlag (kostenfrei)
    Verlag (kostenfrei)
    Resolving-System (kostenfrei)

    Location:
    ZBW - Leibniz-Informationszentrum Wirtschaft, Standort Kiel
    Signature:
    DS 63
    Inter-library loan:
    No inter-library loan

     

    Human-created works represent critical data inputs to artificial intelligence (AI). Strategic behaviour can play a major role for AI training datasets, be it in limiting access to existing works or in deciding which types of new works to create or whether to create new works at all. We examine creators' behavioral change when their works become training data for AI. Specifically, we focus on contributors on Unsplash, a popular stock image platform with about 6 million highquality photos and illustrations. In the summer of 2020, Unsplash launched an AI research program by releasing a dataset of 25,000 images for commercial use. We study contributors' reactions, comparing contributors whose works were included in this dataset to contributors whose works were not included. Our results suggest that treated contributors left the platform at a higher-than-usual rate and substantially slowed down the rate of new uploads. Professional and more successful photographers react stronger than amateurs and less successful photographers. We also show that affected users changed the variety and novelty of contributions to the platform, with long-run implications for the stock of works potentially available for AI training. Taken together, our findings highlight the trade-off between interests of rightsholders and promoting innovation at the technological frontier. We discuss implications for copyright and AI policy.

     
    Export to reference management software   RIS file
      BibTeX file
    Source: Union catalogues
    Language: English
    Media type: Book
    Format: Online
    Other identifier:
    hdl: 10419/300027
    Series: CESifo working papers ; 11099 (2024)
    Subjects: generative artificial intelligence; training data; licensing; copyright; natural experiment
    Scope: 1 Online-Ressource (circa 35 Seiten), Illustrationen
  2. AI Act compact
    Compliance, management & use cases in corporate practice
    Author: Hense, Peter; Mustać, Tea
    Published: 2025
    Publisher:  Fachmedien Recht und Wirtschaft in Deutscher Fachverlag GmbH, Frankfurt

    Export to reference management software   RIS file
      BibTeX file
    Source: Union catalogues
    Language: English
    Media type: Book
    Format: Print
    ISBN: 9783800519606; 3800519607
    Other identifier:
    9783800519606
    Edition: 1. Auflage 2025
    Series: InTeR-Schriftenreihe
    Other subjects: (Produktform)Paperback / softback; (Zielgruppe)The book serves as an introduction and reference work for lawyers and non-lawyers in companies, public authorities and consulting firms; EU AI Act; KI-Verordnung; künstliche Intelligenz; Konformitätsprüfung; Risikoklassen; Datenqualität; Diskriminierungsverbot; Rechenschaftspflicht; Trainingsdaten; Datenschutz; Tea Mustac; Peter Hense; Buch Recht; Bücher für Juristen; Deutsche Gesetze Buch; Deutscher Fachverlag GmbH; Fachmedien Recht und Wirtschaft; Fachmedien Recht und Wirtschaft; Fachbuch Recht; Fachbuch Verlag; Fachbücher Recht; Fachliteratur Recht; Gesetzbuch bestellen; Gesetze Buch; Gesetze Bücher; Gesetze kaufen; Jura Bücher; Jura Recht; juristische Bücher; juristische Literatur; juristischer Fachverlag; Nachschlagewerk; Nachschlagewerk Recht; Recht; Buch Recht und Wirtschaft; Recht und Wirtschaft; Verlag Recht; Verlag Recht; Recht und Wirtschaft; Verlag Recht und Wirtschaft Frankfurt; Compliance Management; Compliance Literatur; Unternehmenspraxis; Use Cases; Rechtshandbuch; CDSMD; DSGVO; DGA; DA; DMA; Data Governance; Risk Management; EU AI Act; AI regulation; artificial intelligence; compliance testing; risk classes; data quality; non-discrimination; accountability; training data; data protection; (VLB-WN)1775: Hardcover, Softcover / Recht/Handelsrecht, Wirtschaftsrecht
    Scope: XIII, 316 Seiten, 21 cm x 14.8 cm, 437 g
  3. Bias and productivity in humans and machines
    Author: Cowgill, Bo
    Published: 8-6-2019
    Publisher:  W.E. Upjohn Institute for Employment Research, Kalamazoo, MI

    Where should better learning technology (such as machine learning or AI) improve decisions? I develop a model of decision-making in which better learning technology is complementary with experimentation. Noisy, inconsistent decision-making introduces... more

    Location:
    ZBW - Leibniz-Informationszentrum Wirtschaft, Standort Kiel
    Signature:
    DS 208
    Inter-library loan:
    No inter-library loan

     

    Where should better learning technology (such as machine learning or AI) improve decisions? I develop a model of decision-making in which better learning technology is complementary with experimentation. Noisy, inconsistent decision-making introduces quasi-experimental variation into training datasets, which complements learning. The model makes heterogeneous predictions about when machine learning algorithms can improve human biases. These algorithms can remove human biases exhibited in historical training data, but only if the human training decisions are sufficiently noisy; otherwise, the algorithms will codify or exacerbate existing biases. Algorithms need only a small amount of noise to correct biases that cause large productivity distortions. As the amount of noise increases, the machine learning can correct both large and increasingly small productivity distortions. The theoretical conditions necessary to completely eliminate bias are extreme and unlikely to appear in real datasets. The model provides theoretical microfoundations for why learning from biased historical datasets may lead to a decrease (if not a full elimination) of bias, as has been documented in several empirical settings. The model makes heterogeneous predictions about the use of human expertise in machine learning. Expert-labeled training datasets may be suboptimal if experts are insufficiently noisy, as prior research suggests. I discuss implications for regulation, labor markets, and business strategy.

     
    Export to reference management software   RIS file
      BibTeX file
    Content information
    Volltext (kostenfrei)
    Volltext (kostenfrei)
    Volltext (kostenfrei)
    Source: Union catalogues
    Language: English
    Media type: Book
    Format: Online
    Other identifier:
    hdl: 10419/202904
    Series: Upjohn Institute working paper ; 19, 309
    Subjects: machine learning; training data; decision algorithm; decision-making; human biases
    Scope: 1 Online-Ressource (circa 31 Seiten)