Securing AI in Healthcare: Prompt Injection to Unauthorized Actions

What happens when an AI-enabled healthcare application encounters malicious instructions hidden inside a clinical record?

Can Azure AI Content Safety Prompt Shields detect malicious instructions in the clinical record, and can the application block processing before generative AI model invocation?

Securing AI in Healthcare: Prompt Injection to Unauthorized Actions is a one-day, in-person instructor-led course designed to investigate these questions through hands-on Microsoft Azure security exercises.

The course applies AI workload security, identity, and trust-boundary principles through a healthcare-specific scenario, emphasizing application-level enforcement, least-privilege authorization, and evidence-based validation of security controls.

Participants examine two critical security boundaries:

• Content Boundary: How malicious instructions embedded in untrusted clinical content are detected, and how application logic enforces the security decision before model invocation.
• Authorization Boundary: How workload permissions govern access to Azure resources, and how controlled READ and WRITE operations provide observable evidence of permitted and restricted actions.

The authorization exercises use direct READ and WRITE tests to evaluate the permissions assigned to a workload identity. They do not demonstrate an AI agent autonomously executing a malicious instruction or attempting the denied WRITE operation.

Through two structured labs, participants move beyond discussing AI security risks to examining what the implemented controls actually demonstrate.

The objective is not simply to understand the threat. It is to test the security boundary, observe the result, and defend the conclusion with technical evidence.

Tuesday, January 12, 2027, 9:00 AM–5:00 PM CST

Class size is limited to 10 participants to preserve the hands-on experience.

The training follows two related security boundaries.

Examine a healthcare AI workflow in which externally sourced clinical content contains an adversarial instruction. Use Azure AI Content Safety Prompt Shields to evaluate the content and examine how application logic enforces the security decision before model invocation.

Examine what an AI-enabled workflow is authorized to do compared with what the business operation actually requires. Evaluate READ versus WRITE requirements and use observable execution evidence to support conclusions about whether access is appropriately constrained.

No Azure subscription is required from participants during the training. Each participant will be assigned an Azure resource group for the hands-on exercises.

Participants encounter the following Microsoft technologies through the healthcare security scenario, instructor-led demonstrations, and structured hands-on exercises:

• Microsoft Azure — Cloud environment supporting the security scenario and lab resources.
• Azure Role-Based Access Control (Azure RBAC) — Authorization controls used to evaluate permitted READ operations and denied WRITE operations.
• Azure Blob Storage — Storage for synthetic clinical documents and controlled authorization tests.
• Azure Functions — Application-level enforcement that blocks processing of detected malicious document content before generative AI model invocation.
• Azure Logic Apps — Workflow orchestration using a managed identity to perform controlled storage operations.
• Azure AI Content Safety Prompt Shields — Detection of indirect prompt-injection indicators in untrusted clinical-document content.
• Microsoft Foundry — Examination of AI-agent behavior when malicious instructions are encountered, distinguishing document instructions, model responses, tool execution, and independently enforced authorization.

Technology exposure varies by module and lab. Listing a technology does not imply that participants independently deploy or configure every service.

• Identifying indirect prompt-injection risk in healthcare AI workflows.
• Validating Prompt Shields detection results.
• Examining application enforcement before model invocation.
• Evaluating authorization requirements against the intended business operation.
• Distinguishing required READ access from unnecessary WRITE capability.
• Using observable technical evidence instead of assumptions to support security conclusions.

• Hands-on experience examining two AI application security boundaries.
• A practical approach for validating security controls through observable evidence.
• Experience connecting prompt-injection detection, application enforcement, and authorization decisions.
• A framework for communicating defensible technical conclusions about AI security risk.
• An NTEKNO Certificate of Completion documenting the training and instructional time.

• The training documents 370 minutes (6 hours 10 minutes) of instructional and hands-on lab time.
• Participants receive an NTEKNO Certificate of Completion and may submit their training documentation to their credentialing organization for consideration.
• CPE or CEU eligibility and any credit awarded are determined by the applicable credentialing organization.

Founder & Chief Learning Architect, NTEKNO, Inc.

Serling A. Davis is a cloud security and technical education professional with 20 years of experience developing and delivering advanced technical training across cybersecurity, cloud infrastructure, telecommunications, and enterprise technologies.

Before founding NTEKNO, Serling served as Principal Content Developer within Microsoft Worldwide Learning's Security, Compliance & Identity organization. He was the principal content developer and global portfolio custodian for AZ-500: Microsoft Azure Security Technologies, responsible for curriculum quality, technical accuracy, practitioner alignment, and continuous program improvement. He also developed the SC-5002 Applied Skills learning path, including GitHub-based technical labs, assessments, validation guidance, and supporting instructional resources.

His career includes technical instructor and curriculum development positions with Palo Alto Networks, Ericsson, Huawei Technologies, Imperva, and Alcatel. His instructional experience spans optical networking, 2G, 3G, 4G LTE and 5G infrastructure, network virtualization, cloud orchestration, and cloud-native security.

Throughout his career, Serling has developed technical training content and delivered instructor-led education for organizations including Verizon Wireless, Vodafone Germany, Swisscom AG, Spark New Zealand, TELUS Canada, Orange SA, China Telecom Hong Kong, Tata Communications Canada, Level 3 Communications, Sprint Corporation, Santee Cooper Power, and the Metropolitan Atlanta Rapid Transit Authority (MARTA).

As Founder and Chief Learning Architect of NTEKNO, Inc., Serling created SecureStack™, a scenario-based cloud and AI security training portfolio incorporating Microsoft technologies such as Defender for Cloud, Sentinel, Purview, Security Copilot, Copilot in Azure, and Azure AI Content Safety Prompt Shields.

His instructional approach emphasizes a central principle: security conclusions must be supported by what technology demonstrably detects, enforces, authorizes, and records—not by assumption.

For Securing AI in Healthcare: Prompt Injection to Unauthorized Actions, Serling applies this methodology through hands-on instruction examining indirect prompt injection, application-layer enforcement, authorization boundaries, and observable technical evidence.

View the complete course details and register directly with NTEKNO.

View course details and registration

Organizations seeking to register employees through an internal procurement or purchase order (PO) process may contact NTEKNO before completing registration.

For organizational registration inquiries, please email [email protected] with the organization name, number of prospective participants, and any applicable purchasing or invoicing requirements.

NTEKNO will review each inquiry and provide information about available registration and payment arrangements.

Important: Submission of a purchase order or organizational inquiry does not constitute confirmed registration or reserve a seat. Participation is subject to availability, applicable payment arrangements, and registration confirmation.

This instructor-led session is limited to a maximum of 10 participants. Registration closes January 7, 2027, at 11:59 PM Central Time.

Security training should demonstrate what works—not simply explain how technology is supposed to work.

NTEKNO SecureStack™ delivers scenario-based, instructor-led training built around realistic security investigations, hands-on technical exercises, and observable evidence.

Led by a former Microsoft Worldwide Learning Principal Content Developer, this course emphasizes the practical distinction between detecting a security threat, enforcing an application-level control, and validating authorization boundaries.

Participants work through structured exercises designed to strengthen technical judgment and support defensible security decisions.

NTEKNO, Inc. is a Texas-based cloud security training company and the developer of SecureStack™, its scenario-based security training portfolio.

SecureStack™ uses realistic security scenarios, structured hands-on exercises, and evidence-based validation to help participants examine how security controls operate in practice.

Training focuses on investigating security conditions, observing control behavior, evaluating technical evidence, and developing defensible conclusions.

NTEKNO offers self-paced, virtual instructor-led (VILT), and in-person instructor-led (ILT) training, with scenarios addressing cloud security, incident readiness, and AI security across industry environments.

Registration: Advance registration is required. Enrollment is limited to 10 participants, and registration closes January 7, 2027, at 11:59 PM Central Time.

Cancellation and refunds: For open-enrollment instructor-led training, cancellation requests submitted at least 26 calendar days before the scheduled course are eligible for a refund of the registration fee or a transfer to a future available open-enrollment session.

Cancellation requests submitted fewer than 26 calendar days before the course are nonrefundable, except where otherwise required by applicable law.

Third-party ticketing and processing fees may be subject to separate refund conditions.

Questions or cancellation requests: Please contact NTEKNO at [email protected] or use Eventbrite's Contact the organizer feature for assistance with registration, cancellations, transfers, or refund requests.

For additional information, visit https://www.ntekno.com.

Info: Course orientation, healthcare AI security scenario, and the evidence-over-assumption approach used throughout the training.

Info: Establish the two security questions that frame the training: what content an AI-enabled workflow should be permitted to consume and what actions the workload should be permitted to perform.

Info: Examine how adversarial instructions embedded within externally sourced clinical content can introduce indirect prompt-injection risk into an otherwise legitimate AI-enabled workflow.

Info: Examine how Azure AI Content Safety Prompt Shields detects adversarial document content and how application logic enforces the resulting security decision before model invocation.

Info: Examine untrusted clinical content, validate Prompt Shields detection evidence, and observe application enforcement before model invocation.

Info: Examine the difference between what an AI-enabled workflow is technically authorized to do and the permissions actually required by the intended business operation.

Info: Examine positive and negative authorization tests to determine whether a required operation succeeds while an unnecessary operation remains constrained.

Info: Examine Microsoft Foundry AI-agent behavior when malicious instructions are encountered in untrusted clinical-document content. Distinguish an agent encountering an instruction from executing a requested operation, and explain why model behavior and natural-language instructions do not establish independently enforced Azure authorization.

Info: Evaluate the authorization boundary by comparing required READ access with unnecessary WRITE capability and use observable execution evidence to determine whether access is appropriately constrained.

Info: Determine what the observed technical evidence supports, identify where that evidence stops, and avoid extending a security conclusion beyond what the tested implementation proves.

Info: Connect prompt-injection detection, application enforcement, and authorization testing to examine how the content and authorization boundaries address different security questions within the same AI-enabled workflow.

Info: Review the technical evidence produced across both security boundaries, connect the findings to the training outcomes, and address final questions.

Event venue & nearby stays

NTEKNO™ SecureStack™ Training | Granite Place at Southlake Town Square, 550 Reserve Street, Southlake, United States

Concerts, fests, parties, meetups — all the happenings, one place.

Get the app →

Discover more events by tag