When Artificial Intelligence Goes Nuts.

The article examines the challenges and risks of artificial intelligence, using OpenAI's early reinforcement-learning experiments and real-world AI applications to illustrate how AI systems can behave unpredictably despite careful programming. Topics include the reward-function problems, emergent an...

Descripción completa

Detalles Bibliográficos
Publicado en:Commentary Vol. 160; no. 2; pp. 10 - 12
Autor principal: MEIGS, JAMES B.
Formato: Artículo
Publicado: Commentary, Inc. Sep2025
Materias:
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=187938012&site=ehost-live
header:
  @attributes:
    shortDbName: ssf
    uiTerm: 187938012
    longDbName: Social Sciences Full Text (H.W. Wilson)
    uiTag: AN
  controlInfo:
    bkinfo:
    jinfo:
      jid:
        00102601
        CMM
      jtl: Commentary
      issn: 00102601
      maglogo: N
    pubinfo:
      dt: Sep2025
      vid: 160
      iid: 2
      pid: 128
      pub: Commentary, Inc.
    artinfo:
      ui: 187938012
      ppf: 10
      ppct: 2
      formats:
        fmt:
          @attributes:
            type: P
            size: 2.1MB
      tig:
        atl: When Artificial Intelligence Goes Nuts.
      aug:
        au: MEIGS, JAMES B.
        affil: Senior fellow at the Manhattan Institute and the former editor ofpopular Mechanics
      su:
        OpenAI Inc.
        Artificial intelligence
        Human-computer interaction
        Reinforcement learning
        Computer programming
      sug:
        subj:
          Artificial intelligence
          Human-computer interaction
          OpenAI Inc.
          Custom Computer Programming Services
          Computer systems design and related services (except video game design and development)
          Other Computer Related Services
          Reinforcement learning
          Computer programming
      ab: The article examines the challenges and risks of artificial intelligence, using OpenAI's early reinforcement-learning experiments and real-world AI applications to illustrate how AI systems can behave unpredictably despite careful programming. Topics include the reward-function problems, emergent and unintended AI behaviors, and the importance of human oversight in AI deployment.
      pubtype: Periodical
      doctype: Article
      src: R
    language: English
    refInfo:
    copyright:
      @attributes:
        flag: N
    holdings:
      @attributes:
        islocal: N