SEC-201 Cybersecurity and governance
Data governance and responsible AI
Before connecting an AI assistant to your documents, three questions: where does your data live, who can query what, and what is the agent allowed to do on its own. In one day, management answers them in writing and leaves with an AI usage charter that applies from the next day.
Objectives
- map the company's data by sensitivity level and hosting location, and decide for each category whether it may be processed by an external AI service, an internal one, or never
- draft a ten-rule AI usage charter for employees, covering the data that may be shared, the approved tools, the checking of results and accountability
- set the governance rules of an agent: actions allowed on its own, actions subject to human validation, logging, a named owner, response time in case of an incident
- apply personal data protection obligations to an AI project (Ivorian regulation, GDPR for European customers): purpose, information of individuals, retention period, processors, hosting
- require, from a project affecting people (recruitment, appraisal, credit granting), a bias check, an explanation of decisions and human validation, and read a bias report to reach a decision
Audience
Executives, IT directors and IT managers, HR directors, legal and compliance departments, data protection officers, quality and risk managers.
Prerequisites
Hold a responsibility over the company's data, systems or people. No technical skills. Technical requirements: a computer or tablet to consult the materials; for remote delivery, a stable connection, microphone and camera; participants bring the list of AI tools already used in the company, official or not, and an AI project under way or planned.
Programme
- 40% inputs, 60% decision workshops
- Mapping of the company's data during the session
- Drafting of the charter in small groups, then consolidation
- Case studies inspired by company situations (assisted screening of applications, pre-assessment of credit files, a sales agent with access to the customer file), including the reading of a bias report and of a decision explanation
- Decision game on an incident (data leaking to a public assistant)
- Materials handed out: a data classification grid, a usage charter template, an agent governance scope template, a compliance checklist, a jargon-free glossary
- Final deliverable: the company's charter
- Follow-up point at three weeks
Format
In person preferably; in-company (extended management committee) or open enrolment (inter-company); 6 to 12 participants; 7 hours, 1 day; remotely, 2 sessions of 3 h 30 one week apart, the charter being drafted in between.
Assessment
Beforehand, a positioning questionnaire (tools in use, maturity). During the course, a decision case study corrected as a group. At the end, the charter and the governance grid are presented and scored on a grid, an immediate feedback survey and a named certificate of course completion. At three weeks, a follow-up point on the circulation of the charter.
Ask for this course
A person answers you and sends you a quote.
Related courses
- Cybersecurity and governanceIncident management and continuity planWho does what in the first hour, how to resume operations, how to communicate.SEC-2031 day
- Cybersecurity and governanceEssential cybersecurityMost incidents start with a click: a fake email, a pressing WhatsApp message, a reused password.SEC-1012 days
- Cybersecurity and governancePersonal data protection: Ivorian regulation and GDPRObligations, register of processing activities, rights of individuals, processors and transfers, applied to your customer and HR files.SEC-1021 day