Two linked cybersecurity developments are drawing attention to the security risks around AI tooling and software supply chains. India’s national cyber agency, CERT-In, has warned that an AI-assisted threat actor known as Crafty Falcon is actively targeting organisations in the banking, financial services and insurance sector. Separately, researchers reported that a backdoored Docker image associated with ARTEX AI exposed 62 internet-facing consoles to potential root-level compromise.

What happened

According to CERT-In, the threat activity is active and ongoing and affects organisations in India’s BFSI sector. The agency said the attacker is using publicly available AI-driven offensive security tools. Based on the source material, CERT-In’s warning is a confirmed alert from a national incident response authority, but the operational details available in the source are limited.

In a separate but related security issue, researchers identified a compromised Docker image tied to ARTEX AI. The report said the backdoored image exposed 62 consoles and could allow root access on affected systems. That points to a software supply-chain problem: businesses may deploy what appears to be a legitimate AI-related component, only to introduce malicious code directly into their own infrastructure.

Why it matters for European businesses

Although the CERT-In alert is focused on India, the underlying risks are relevant well beyond one market. European businesses are increasingly experimenting with AI tools, self-hosted models, containerised applications and third-party automation services. That creates a broader attack surface for threat actors using AI-assisted methods or targeting AI infrastructure.

The main business concern is not only that attackers may use AI to scale phishing, reconnaissance or exploitation, but also that organisations may introduce risk through the way they adopt AI tools. A malicious or tampered container image can provide attackers with privileged access, potentially affecting internal systems, customer data, development environments and cloud workloads.

For SMEs, the practical issue is clear: AI adoption can create security exposure if teams deploy tools quickly without the same review applied to other production software. This is especially important where companies use open-source packages, Docker images, hosted admin panels or AI components connected to sensitive data and business workflows.

Who may be affected

• Financial services, fintech and insurance firms using AI tools in operations, analytics or customer service
• SMEs and digital teams deploying containerised applications without strong image verification processes
• E-commerce and online businesses running internet-facing management consoles or cloud workloads
• IT and DevOps teams responsible for infrastructure, CI/CD pipelines and software deployment
• Regulated businesses that must protect customer data and demonstrate stronger security controls

What companies should consider

• Review AI-related software sources: Verify where Docker images, plugins and model-serving components come from before deployment.
• Harden container security: Limit privileged access, restrict root usage where possible, and review exposed management consoles.
• Check for internet-facing admin interfaces: Consoles exposed to the public internet increase the impact of any backdoor or weak authentication issue.
• Strengthen supply-chain controls: Use trusted registries, image signing, vulnerability scanning and approval workflows for production deployments.
• Treat AI tools like other business-critical software: Apply procurement, security review and patching processes rather than allowing informal adoption.
• Monitor for unusual activity: Look for unauthorised container behaviour, privilege escalation and suspicious outbound connections.
• Prepare incident response steps: If a suspicious image has been deployed, isolate affected systems, rotate credentials and investigate for persistence.

For European companies, the broader lesson is that AI-related cyber risk now extends beyond model misuse into infrastructure compromise and software supply-chain exposure. Businesses adopting AI should ensure security controls keep pace with deployment speed.