GoDaddy and its subsidiary 123 Reg experienced an AI-driven cyberattack on September 18, 2026, where attackers used the agentic AI tool ARTEX to exploit API vulnerabilities and harvest customer data.
123 Reg, a UK domain registrar owned by GoDaddy, reported unauthorized access to customer data, including names, phone numbers, and business tax registration numbers. A spokesperson said that the company observed the ARTEX AI security-testing tool during the intrusion. The registrar said access was limited to reading the data and that the intruder may have made copies.
Risk classification
- Primary risk domain: 4 Malicious actors
- Primary risk subdomain: 4.2 Cyberattacks, weapon development or use, and mass harm
The incident involved malicious actors using an AI-driven penetration testing tool (ARTEX) to conduct a cyberattack against GoDaddy and 123 Reg systems.
Additional risk subdomains
- 2.1 Compromise of privacy by obtaining, leaking or correctly inferring sensitive information: The AI-driven attack successfully read and potentially copied non-public customer relationship data, settings, and VAT numbers.
- 4.3 Fraud, scams, and targeted manipulation: The harvested customer data directly enables highly convincing, targeted phishing and impersonation scams pretending to be GoDaddy support.
Causal factors
- Entity: AI
- Intent: Intentional
- Timing: Post-deployment
The cyberattack was executed by the AI-driven agentic tool ARTEX, which was intentionally deployed by malicious actors to exploit GoDaddy's APIs post-deployment.
EU AI Act risk tier
High Risk: The AI system is an offensive cyber-warfare/penetration testing tool used to automate attacks on digital infrastructure, posing significant implications for public security and data protection.
AI system and alleged parties
- AI system: ARTEX
- AI purpose: Target Identification; Code Generation
- Behaviour type: Agent
- Alleged developer: ARTEX developers, AI agent system developers
- Alleged deployer: Threat actors, hackers, Cybercriminals, AI agent system deployers, Agentic threat actors
- Alleged harmed parties: Privacy, Information security, GoDaddy sites, GoDaddy, Domain registrars, Customers, 123 Reg customers, 123 Reg
Harm severity
Severity ratings are based on this scale.
Highest direct severity in any category: Substantial. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.
- Physical: direct Negligible, indirect Negligible
- Infrastructure: direct Negligible, indirect Negligible
- Property: direct Negligible, indirect Negligible
- Financial: direct Negligible, indirect Negligible
- Environmental: direct Negligible, indirect Negligible
- Malicious content: direct Negligible, indirect Negligible
- Differential treatment: direct Negligible, indirect Negligible
- Civil rights: direct Negligible, indirect Negligible
- Democracy: direct Negligible, indirect Negligible
- Privacy: direct Substantial, indirect Negligible
- Psychological: direct Negligible, indirect Negligible
- Epistemic: direct Negligible, indirect Negligible
- Child sexual exploitation and abuse: direct Negligible, indirect Negligible
Privacy
Reported: The report explicitly describes privacy violations where customer profile, settings, and contact information were read and potentially copied.
Directly caused: The AI-driven attack successfully read customer profile and settings data, including published contact details, VAT numbers, and account preferences from GoDaddy and 123 Reg APIs.
Inferred additional harm: It is highly likely that the attackers created permanent copies of the exposed customer data, resulting in ongoing privacy exposure for thousands of website owners.
People affected
- Occurrences reported: 2
- People reportedly exposed: 10000
Potential causes
Management
- Legacy Security Deficiencies: GoDaddy historically struggled with MFA implementation and threat monitoring.
Technology
- Agentic AI Pentesting Tools: Open-source tool ARTEX enabled automated, rapid discovery of API weaknesses.
- Abuse of Open-Source AI Skills: Attackers cloned Anthropic's open-source skills to automate exploitation.
- Vulnerable API Interfaces: APIs exposed structured customer profile data to unauthorized queries.
Human Factors
- Lack of Multi-Factor Authentication: Lack of 2SV leaves users vulnerable to post-attack phishing scams.
- Susceptibility to Phishing Lures: Exposed profile details make targeted phishing emails highly convincing.
Process and Methods
- Inadequate API Rate Limiting: The API allowed rapid, automated reading of thousands of customer records.
- Delayed Attack Detection: The intrusion occurred on Sept 18 but was not detected until Sept 19.
Regulatory Environment
- Dual-Use AI Software Release: Open-source release of agentic tools like ARTEX lacked safety controls.
Information quality
- Classification confidence: High
- Reason for confidence: The reports provide consistent, detailed accounts of the cyberattack from GoDaddy and 123 Reg notifications, including the specific date, the AI tool used (ARTEX), and the nature of the data exposed. While the exact number of affected customers is not disclosed, the technical mechanism and downstream risks are clearly explained.
- Ambiguities identified: The exact number of affected customers is not specified; GoDaddy has not officially confirmed if the W+M and 123 Reg attacks were part of the same coordinated operation, though they are highly likely to be linked.
- Alternative interpretations: The incident could be viewed purely as a traditional data breach, but the explicit involvement of agentic AI tools to automate and accelerate the API exploitation justifies its classification as an AI safety/security incident.
An AI-driven cyberattack leveraged the Chinese-developed agentic tool ARTEX to exploit API vulnerabilities and harvest customer data from GoDaddy and 123 Reg. While the direct national security impact is minor, the incident highlights an evolving threat where agentic AI dramatically lowers the cost and increases the speed of automated cyber operations at scale.
- Overall national security impact: Minor
- Scope: Multiple nations
- Primary target: United States and United Kingdom
- Alleged perpetrator: Unknown
Threat characteristics
- Imminence: Long-term. The active incident has been contained, representing an ongoing strategic concern regarding automated vulnerability exploitation.
- Autonomy: Human-supervised. The agentic AI tool autonomously mapped targets and adapted code to exploit vulnerabilities, directed by human operators.
- Novelty: Evolved capability. Represents a significant evolution of cyber threats, leveraging agentic AI to automate and accelerate API vulnerability exploitation.
Impact by dimension
- Physical security: Negligible. No physical infrastructure, kinetic targeting, or critical systems were compromised or affected by this data breach.
- Information security: Minor. Stolen customer profile and contact data increases downstream phishing risks, but did not compromise classified networks or intelligence capabilities.
- Sovereignty: Negligible. The breach targeted commercial web hosts and did not impact state authority, electoral systems, or core government operations.
- Economic security: Minor. Affected commercial domain registrars but did not disrupt financial systems, steal strategic technologies, or cause critical supply chain failures.
- Societal stability: Negligible. No evidence of civil liberty violations, algorithmic oppression, or threats to societal stability.