30 Days Gen AI Risk Trial -Start Now
Skip to main content
Policy in practice · Practical playbook

Resolve the risks of a shared AI account

A shared workspace and a shared login create different operating questions. Before approving either, establish who can see conversations, who can change access and how actions are attributed. If several people use one identity, the organization may struggle to separate business continuity from individual access removal.

For Application owners, team managers and security administrators

Synthetic example

A department uses one assistant login

A small operations team has been using a single account for drafting. New members can see old conversations, while nobody is sure who owns the account or its recovery process.

What you are working with

  • The people using the identity and its business owner.
  • The information visible in existing conversations and projects.
  • The provider's available account and team-access arrangements.

A safer approach

  • Separate a collaborative workspace from shared credentials.
  • Have the owner review access and sensitive conversation exposure.
  • Plan a move to an approved arrangement with identifiable users.

Expected outcome: The team has a documented access model and a practical migration path, with ownership and offboarding responsibilities no longer depending on a shared secret.

Put it into practice

Work through the procedure

  1. Identify the form of sharing

    Ask whether people share credentials, belong to a managed workspace or use publicly shared conversation links. Map who can see and change what. Do not assume that a collaboration feature exposes the same information as several employees signing in through one identity.

  2. Review the existing exposure

    Have an authorized owner inspect the types of business information accessible through the arrangement. Avoid distributing conversation contents during the review. If inappropriate access or sensitive submissions are suspected, route them through the incident process rather than treating account cleanup as a complete response.

  3. Choose an accountable working arrangement

    Check the provider's current supported options and the organization's procurement requirements. Prefer a reviewed setup that can identify users and remove their access independently. Define a temporary boundary for current work while access and necessary business records are transferred through authorized methods.

  4. Verify ownership and member changes

    Test how the selected arrangement handles a new team member and a departure using harmless content. Confirm who controls administration and account recovery. Update onboarding instructions and review any shared links or old credentials separately; changing the login model does not settle every exposure.

Evidence before approval

What to check before proceeding

1. The access model is understood

Ready when
Owners can explain the difference between membership, shared credentials and shared content.
If the check fails
Map those relationships before granting additional team access.

2. Individual access can be managed

Ready when
The reviewed arrangement supports the organization's accountability and offboarding needs.
If the check fails
Choose a different approved route or limit the task until the gap is resolved.

3. Existing content has an appropriate audience

Ready when
The owner has reviewed shared business records and access permissions.
If the check fails
Restrict access as appropriate and assess any suspected exposure separately.

Common mistakes to avoid

  • Calling a shared password a team workspace because several employees can log in.
  • Assuming endpoint activity identifies which person performed an action inside a provider account used by several people.
Workforce AI Security

Evaluate this workflow with Aona

Where Aona can help

Aona's supported endpoint visibility and data policies can contribute information about AI use on deployed devices. They can support the team's chosen input restrictions while account ownership is resolved.

What to confirm

Aona does not make a shared provider identity individually accountable or control that provider's conversation sharing. Verify account membership, credentials and access through the relevant provider administrator.

Turning this policy into an operational rollout?

Discuss the teams, devices and AI tools in scope, who will own the policy, and which deployment and evidence requirements need to be met before rollout.

FAQ

Questions about this workflow

No. A reviewed workspace may provide appropriate collaboration and administration. Establish its actual permissions and account arrangement. The concern is unclear access or accountability, not the word shared on its own.
Technical evaluation

Turning this policy into an operational rollout?

Discuss the teams, devices and AI tools in scope, who will own the policy, and which deployment and evidence requirements need to be met before rollout.

Govern Shared AI Accounts at Work | Aona AI