ServiceNow AI Platform EUVD-2026-86393

Exploitation requires a valid authenticated account on the ServiceNow AI Platform (PR:L) - any low-privilege user session suffices; no elevated role, no user interaction, and no special client configuration are needed, and the flaw is reachable over the network against default configurations. The bounding factor is authentication: an attacker must first obtain or possess legitimate platform credentials (e.g., an internal user, a compromised account, or a low-tier tenant user). Hosted instances have already been updated by ServiceNow; self-hosted and partner deployments remain exposed until the KB3159623 update/patched release is applied, so unpatched self-managed installations are the primary realistic risk surface.

This is a credible priority for ServiceNow customers, though not an emergency. The vendor-supplied CVSS 4.0 score of 8.4 (AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N/SC:H/SI:H/SA:N) describes a network-reachable, low-complexity authorization bypass exploitable by any authenticated low-privilege user with no user interaction. The high confidentiality impact on the vulnerable system (VC:H) plus high confidentiality AND integrity impact on subsequent systems (SC:H/SI:H) reflects the description's warning that unauthorized data access could enable 'further unintended access' - i.e., a pivot beyond the initial data exposure. Mitigating the urgency: there is NO EPSS score, NO CISA KEV listing, and NO public PoC in the provided data, and ServiceNow explicitly states it is 'not currently aware of malicious exploitation.' Exploitation status is therefore 'no public exploit identified at time of analysis.' The requirement for a valid authenticated account (PR:L) meaningfully raises the bar versus a pre-auth flaw. Net: real, patch-worthy, and rated High/Critical by the vendor, but current signals point to no active exploitation - treat as a prompt-patch item rather than a fire-drill. Note the CWE is listed as N/A; the flaw is an authorization (access-control) bypass, not a true authentication bypass despite the tag.

An attacker who holds any legitimate low-privilege account on a ServiceNow AI Platform instance (or a self-hosted deployment) sends normal authenticated requests that the platform fails to properly authorize, returning records, fields, or objects the account is not entitled to see. Because the access-control check is missing rather than merely weak, the attacker can enumerate and read other tenants' or users' data and potentially leverage that exposed information to reach further restricted functionality. No public exploit code is identified at time of analysis, and no active exploitation is known.