Tillbaka till jobb

Legal and Information Security Analysis of AI Services – Team Assignment

CybersecurityPrincipal
Stockholm, SwedenContract · On-site Publicerad 17 november 2026 Ansök senast 23 november 2026
Språk: SweMinsta erfarenhet: 9 år

Liam Tran

HR Administrator

cv@veritaz.se

Efterfrågade kompetenser

Identity and Access ManagementInformation SecurityCloud SecurityRetrieval-Augmented GenerationModel Context ProtocolTechnical ArchitectureAPI securityGenerative AI Security

Om tjänsten

Role Summary

A Swedish public authority requires a consulting team of up to three specialists to conduct coordinated legal, information security, and technical risk and impact assessments of ChatGPT and Copilot. The objective is to produce a consolidated decision-making basis for the potential use of these generative AI services by the authority’s employees.

The assignment includes mapping relevant service versions, licenses, and functionality; defining use cases and associated information handling; assessing legal, information security, and technical risks; and recommending safeguards, configurations, conditions, and usage restrictions. The analysis will cover ordinary chat without internal documents as well as the upload and analysis of documents processed by the AI services. Subject to available time and joint prioritization, the scope may also include AI integrations with internal systems through MCP or APIs.

The proposed team must collectively cover legal expertise in data protection, confidentiality, AI regulation, supplier terms, and international data transfers, together with technical expertise in cloud services, generative AI architecture, data flows, access control, integrations, and security risk assessment. The legal and technical workstreams must be coordinated into one coherent delivery.

Key Responsibilities

  • Map relevant ChatGPT and Copilot versions, licenses, features, technical characteristics, and material differences between service offerings.
  • Work with the client to define applicable use cases, information categories, processing activities, and required internal participation.
  • Conduct legal assessments covering data protection, confidentiality, supplier terms, international data transfers, and the applicability of AI-related regulation.
  • Perform data protection impact assessments where required and document legal risks, conditions, and residual risks.
  • Map how prompts, uploaded documents, logs, generated content, and other information flow through the services and their supporting infrastructure.
  • Assess information security and technical risks associated with cloud delivery, storage, deletion, training-data use, tenant isolation, permissions, APIs, and integrations.
  • Evaluate technical architecture and safeguards for generative AI services, including identity and access controls, configurations, data handling, and integration security.
  • Recommend proportionate safeguards, configurations, contractual conditions, restrictions, and governance measures for each identified use case.
  • Coordinate legal and technical findings and produce a consolidated analysis report for each service, clearly describing differences between versions and licenses.
  • Prepare a comparative summary of ChatGPT and Copilot, recommendations for use, residual-risk documentation, and decision material for designated management groups.
  • Deliver an oral presentation and structured handover of results in Swedish.
  • Participate in a joint project launch and ongoing checkpoints, including approximately three workshops, review meetings, or presentations.
  • If prioritized within the agreed schedule and budget, assess AI connections to internal systems through MCP or APIs, including agent and tool-integration risks.
  • Provide a delivery plan covering team composition, responsibilities, estimated hours by consultant and activity, analytical methods, quality assurance, schedule, and dependencies on client participation.
  • Appoint an assignment lead responsible for coordination, schedule management, quality assurance, and primary contact with the client.

Technical Stack & Requirements

  • AI services: Practical knowledge of ChatGPT and Microsoft Copilot, including service editions, licensing models, enterprise functionality, and information-handling characteristics.
  • Generative AI architecture: Documented practical experience assessing generative AI services, including APIs, integrations, permissions, data flows, and security controls.
  • Cloud and information security: Demonstrated ability to assess cloud-service architecture, information processing, technical threats, residual risks, and appropriate safeguards.
  • Data lifecycle: Experience analyzing the storage and deletion of prompts and logs, use of customer data for model training, tenant isolation, and handling of AI-generated material.
  • Integration security: Experience with APIs and preferably AI agents, MCP-based tool integrations, and their associated authorization and data-exposure risks.
  • RAG: Advantageous experience evaluating retrieval-augmented generation solutions and their data-handling and security implications.
  • Information classification: Advantageous experience applying recognized public-sector information-classification methods and translating classifications into risk assessments and safeguards.
  • Assessment methods: Documented competence in information security and risk assessment using ISO 27001/27005 or an equivalent method.

Qualifications & Experience

The proposed team may consist of one to three consultants. The required capabilities may be held by one consultant or distributed across the team, but all individuals cited to meet mandatory requirements must participate actively in the assignment. A named specialist must be responsible for each of the legal and information security/technical workstreams, and one named consultant must act as the overall assignment lead. The same individual may hold multiple responsibilities.

The specialists responsible for the legal and information security/technical assessments must meet competence level 4 under the applicable framework agreement. This represents very high subject-matter expertise, experience delivering large and highly complex assignments to a very high standard, strong independence, and the ability to lead or take responsibility for major engagements. This level is normally reached after approximately 9–12 years of relevant professional experience.

The team must demonstrate documented competence in data protection and impact assessments under GDPR; confidentiality assessments for cloud-service use; review of supplier terms and international data transfers; information security and risk assessment using ISO 27001/27005 or an equivalent method; cloud services, data-flow analysis, and technical safeguards; and knowledge of the EU AI Act and its application to AI services. At least one consultant must have practical experience in the technical architecture and security assessment of generative AI services, including APIs, integrations, access control, and data flows.

The supplier must provide at least two completed reference assignments that collectively demonstrate legal analysis of cloud or AI services, information security or technical risk assessment, and production of reports, recommendations, or executive decision material. At least one reference assignment must have been completed for a public authority. For every submitted reference, at least one consultant proposed for this engagement must have participated, with their role, responsibilities, work performed, and deliverables clearly described.

Additional merit will be given for knowledge of Swedish archival legislation and the principle of public access as they relate to prompts, logs, and AI-generated material; experience evaluating data retention, deletion, model-training use, RAG, and tenant isolation; experience with public-sector information classification; experience presenting combined legal and information security decision material to authority management; and practical knowledge of AI agents, MCP, and tool integrations.

Work Logistics & Location

This is a temporary consulting assignment for a Swedish public authority. The desired start is immediate, and the expected duration is approximately two months. The exact work location and remote or on-site arrangement are not specified. The engagement is expected to include approximately three joint workshops, checkpoints, or presentations: one focused on ChatGPT, one on Copilot, and one consolidated session covering results and outstanding questions. The final number and format of meetings will be agreed with the client.

Reports and decision material must be delivered in Swedish, in an editable format, and follow a common template approved by the client. Any additional services or analyses outside the defined scope require a separate written agreement within the applicable contractual framework.

Intresserad av denna tjänst?

Skicka din ansökan så återkommer vi inom kort.

Ansök nu
Legal and Information Security Analysis of AI Services – Team Assignment | Veritaz