Reinventing AI: Is It the Time for a New Paradigm?

The article proposes a paradigm shift in artificial intelligence (AI) methodologies, moving intelligence from centralized cloud systems to billions of small devices with onboard central processing units (CPUs), reflecting a potential return to the era of distributed computing. By embedding learning...

Descripción completa

Detalles Bibliográficos
Publicado en:Communications of the ACM Vol. 68; no. 11; pp. 37 - 41
Autor principal: Gori, Marco
Formato: Artículo
Publicado: Association for Computing Machinery Nov2025
Materias:
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=hlh&AN=188947690&site=ehost-live
header:
  @attributes:
    shortDbName: hlh
    uiTerm: 188947690
    longDbName: Humanities International Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    jinfo:
      jid:
        00010782
        ACM
      jtl: Communications of the ACM
      issn: 00010782
      maglogo: N
    pubinfo:
      dt: Nov2025
      vid: 68
      iid: 11
      pid: 68
      pub: Association for Computing Machinery
    artinfo:
      ui:
        188947690
        10.1145/3731676
      ppf: 37
      ppct: 4
      formats:
      tig:
        atl: Reinventing AI: Is It the Time for a New Paradigm?
      aug:
        au: Gori, Marco
        affil: University of Siena, DIISM, Siena, Italy, Italy
      su:
        Artificial intelligence
        Central processing units
        Distributed computing
        Language models
        Machine learning
        Cognitive robotics
      sug:
        subj:
          Artificial intelligence
          Central processing units
          Distributed computing
          Language models
          Machine learning
          Cognitive robotics
      ab: The article proposes a paradigm shift in artificial intelligence (AI) methodologies, moving intelligence from centralized cloud systems to billions of small devices with onboard central processing units (CPUs), reflecting a potential return to the era of distributed computing. By embedding learning processes directly within these devices, AI systems could interact continuously with their environments, allowing robots and agents to develop cognitive abilities through experiential engagement, akin to developmental robotics in nature. Time becomes a central factor, as learning and evaluation occur simultaneously, enabling adaptive and context-sensitive intelligence rather than reliance on static training-test separations. Integrating large language models and meta-learning mechanisms with environmental interaction frameworks further enhances the potential for scalable, actionable, and socially distributed AI systems.
      pubtype: Periodical
      doctype: Article
      src: R
    language: English
    refInfo:
    copyright:
      @attributes:
        flag: Y
      dt:
        @attributes:
          year: 2025
    holdings:
      @attributes:
        islocal: N