GCRI · Yapay Zekâ ve Otonom Sistemler

Ajan sistemler

Agentic AI sistemleri pursue goals by planning and taking actions: calling tools, browsing, writing and running code, or operating other software, sometimes over long tasks with little supervision. Autonomy brings usefulness and new failure modes, from compounding errors to misuse of granted permissions. This page sets out how agents fail, how their actions can be bounded and overseen, and what evidence serves those who build, deploy and govern them.

01Sicil

Agentic systems on the record.

02Soru

Agentic systems: how the risk develops.

Tool access, delegated authority and safe execution. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.

Mekanizma

Errors compound over steps

Each step in a long task can go wrong, and an early mistake shapes later actions. Small per-step error rates add up over many steps.

Mekanizma

Permissions define the blast radius

An agent can only do harm with the access it has been given. Broad credentials, write access and payment rights turn errors or manipulation into real-world consequences.

Mekanizma

Untrusted inputs steer behaviour

Agents read web pages, documents and messages that may contain instructions. Prompt injection can redirect an agent to act against its user.

Mekanizma

Oversight must be designed in

Humans cannot review every action of a fast agent. Approval steps for consequential actions, logging and the ability to stop and roll back are design choices.

03İzleme

Agent oversight console.

Bir ekibin bu konuyu izlemek için kullanacağı ekran: ölçütleri, durumları ve kararları. Adlar örnektir ve değerler açıklayıcıdır.

Agent oversight console

Agentic systems · Agent sandbox · Session AG-6627

Human approval pendingGCRI · Yapay Zekâ ve Otonom Sistemler

Tool calls

  1. 10:34Bilgisearch(documents) · read-only
  2. 10:51Bilgirun_model(flood_nowcast) · sandbox
  3. 10:51Geçitsend_message(external) · needs approval
  4. 10:53Errorwrite_file(outside workspace) · denied

Politika

Kapsam
workspace only
External actions
human approval
Spend limit
43 units
Kill switch
Armed

Actions by approval class

050100count this weekOtomatikKaydedildiOnaylandıDeniedÜst kademeye iletildi
Agent oversight console: the screen’s structure, terms and states for Agentic systems.Örnek adlar ve örnekleyici değerler; bir tasarımdır, kayıt veya canlı bir sistem değildir.
04Alanlar genelinde

Agents that act need guardrails in every domain they touch.

Agentic system studies are designed for the Üstel Teknoloji serisi. The OWASP Top 10 for LLM applications lists failure modes such as excessive agency and prompt injection; the page describes them without exploit detail.

  • Incident reports involving agents.
  • Designs for oversight and permission scoping.
05Yöntemler ve kanıt

How agentic systems is modelled, measured and tested.

Model, varsayımları, verileri ve belirsizliği, bir bilim insanının veya mühendisin kontrol edeceği biçimde. Yapılar yayımlanmış yöntemleri izler. Çizilen değerler bulgu değil, açıklayıcıdır.

Sistem modeli

The couplings between the systems in agentic systems: signs of influence, a reinforcing and a balancing loop, and cross-links where stress jumps.

++−+Yapay zekâVeriTopluluklarEkonomiR pekiştiriciB dengeleyici
Causal loop diagram from the page’s systems; signs show the direction of influence. Illustrative structure for a system dynamics or network model, not estimated couplings.
FAIR yol gösterici ilkeleri · GO FAIR ve FORCE11W3C PROV · World Wide Web ConsortiumBelirsizliklerin tutarlı ele alınmasına ilişkin kılavuz · IPCCCEOS Analysis Ready Data · Yer Gözlem Uyduları Komitesi
06Göstergeler

Neyin ve ne kadar iyi ölçüldüğü.

Her gösterge birimini, sıklığını, ardındaki kaynak türünü ve bir değerin taşıyacağı kanıt etiketini taşır. Hiçbir değer gösterilmez: değerleri bir program kendi yetkisi altında sağlar.

GöstergeBirimSıklıkKaynak türüEtiket
Task success rate on agent benchmarkspercent of tasks completed correctlyper evaluationAgent benchmark resultsBİLDİRİLMİŞ
Actions requiring human approvalpercent of consequential actionsper deploymentDeployment configurationBEYAN EDİLDİ
Prompt injection success ratepercent of attempts succeedingper evaluationSecurity testingGÖZLENEN
Agent incidentscount by severityaylıkIncident reportsBİLDİRİLMİŞ
Actions logged and attributablepercent of agent actionsper deploymentAudit logsGÖZLENEN
Mean task length before human interventionsteps or minutesper evaluationEvaluation logsGÖZLENEN
07Kaynaklar

Kanıtın nereden geldiği.

Her biri lisans sınıfıyla birlikte, adı belirtilmiş kamuya açık veri kümeleri ve kurumsal kayıtlar. Anlaşmaya dayalı kaynaklar yalnızca koşullarına göre kullanılır ve asla yeniden yayımlanmaz.

KaynakSağladıklarıLisans sınıfı
OWASP Top 10 for LLM Applications and agentic security guidanceRisks such as prompt injection and excessive agencyOpen, CC BY-SA 4.0
Published agent benchmarks such as SWE-bench and related suitesTask-based evaluation of agentsOpen, licence per benchmark (to check)
NIST Yapay Zekâ Risk Yönetimi Çerçevesi (NIST AI Risk Management Framework)Practices for governing and measuring AI riskPublic domain, US government
AI Incident DatabaseReported incidents involving AI sistemleriOpen, licence to check
Developer system cards describing agent safeguardsDeveloper-reported agent evaluations and controlsPublic documents, terms per developer
08Kararlar

Kim karar verir ve kanıt neye hizmet eder.

Yapay Zekâ ve Otonom Sistem prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.

Kim karar verirSoruBu sayfanın hizmet ettiği kanıt
Deploying organisationsWhich actions an agent may take alone and which need approvalPermission patterns; evaluation results
Security teamsHow to protect agents from manipulationInjection testing; incident evidence
DüzenleyicilerWhat oversight to require for autonomous AI actionsOversight design evidence

Bu konunun ürettiği çıktılar

  • Permission and approval design patterns for agents
  • Agent evaluation protocols covering long tasks
  • Prompt injection test suites for deployments
  • Logging and rollback requirements
09Temel sorular

What designers and evaluators ask.

  • What actions can the agent take, with whose authority, and which are reversible?
  • How does it behave when instructions are ambiguous, conflicting or malicious?
  • Can a human see what it is doing, understand why and stop it in time?
  • How do errors compound over long sequences of steps?
10Mekanizmalar

How agents fail.

An agent combines a model’s reasoning with permissions in the world. Small errors in early steps can compound across a long task. Agents can misinterpret goals, pursue them in unintended ways or take shortcuts that satisfy a metric but not the intent. Content the agent reads, such as a web page or document, can contain instructions designed to hijack it, an attack known as prompt injection.

The more permissions an agent holds, the larger the potential harm. Sistem that coordinate several agents add further failure modes, as agents pass errors to each other or interact in unforeseen ways.

11Yöntemler ve veriler

Designing for safe execution.

  1. 01

    Least privilege

    Granting only the permissions a task needs, with sandboxes and spending or action limits, drawing on siber güvenlik ve kriptografi.

  2. 02

    Human oversight points

    Requiring confirmation for consequential or irreversible actions. Too many prompts lead to rubber-stamping, so placement matters.

  3. 03

    Task-based evaluation

    Testing agents on realistic multi-step tasks in controlled environments, using Yapay zekâ değerlendirme ve güvence mühendisliği. Simulated environments miss some real-world complexity.

  4. 04

    Logging and audit

    Recording actions and reasoning traces so incidents can be reconstructed. Traces may not fully reflect how a model reached a decision.

13Where this page’s references

Where this page’s references lead.

Bu sayfanın karşılaştırıldığı kuruluşlar ve belgeler. Durumlar ve sürümler değişir; her bağlantı yayımlayan kuruluşa gider.

14Sorular

Sorular ve yanıtlar.

What is an AI agent?
An AI system that pursues goals by planning and taking actions, such as using tools, browsing or running code, with some degree of autonomy.
What is prompt injection?
An attack in which instructions hidden in content the AI reads, such as a web page or document, cause it to act against its user’s intent.
How can agent risk be reduced?
By limiting permissions to what the task needs, requiring approval for consequential actions, logging all actions, testing against manipulation and keeping the ability to stop and undo.
Should agents ever act without human approval?
For low-consequence, reversible actions within tight limits, often yes. For consequential or irreversible actions, the design question is where and how a human confirms.
Can agents be evaluated before deployment?
Partly, in controlled environments with realistic tasks. Monitoring after deployment remains necessary because real environments are more varied.
Does the platform build agents?
Its work is evaluation and method. Any build work for Consortium projects runs through Foundry under its own governance.
RaporlarKatkıda bulunun

Bu sayfadan kanıt katkısında bulunun.

Raporlar is the Consortium’s quarterly publication. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work.

GCRIYapay Zekâ ve Otonom SistemlerÜstel TeknolojiLabs Kanıtı
Yapay Zekâ ve Otonom Sistemler

Bir yöntem, bir veri seti ya da bir soru getirin.

Bilgi ve başvurular bugün açıktır. Bir başvuru bir referans numarası alır ve bir sonraki adımı belirtir; başvurular, ödemeler, işlem kapasitesi tahsisleri ve kayıtlar yalnızca ekipleri, koşulları ve inceleme kapasiteleri mevcut olduğunda açılır.

The Global Centre for Risk and Innovation

Science · evidence · measurement

Not for profit

Our articles, maps and twins are free to read and free to republish. Readers and partners keep it that way.

Partner with us

Who funds us

GCRI is a not-for-profit institution. To ask who funds a piece of work, use the funding questions form.

Independence

Editors and the expert reviewers named on each article decide what we publish. Funders and partners do not review or approve articles before publication.

Licence

Articles: CC BY 4.0. Figures, maps, twins and data: CC BY 4.0. Other uses: ask through the permissions form.