Workshops: EU AI Act & Compliance
Shadow AI: How to Detect and Govern It – Workshop
Shadow AI happens when employees use AI tools without approval, usually on personal accounts and with good intentions. This workshop shows IT, security and privacy teams how to make that usage visible, assess the risk realistically and channel it into secure, approved options.
Who is this workshop for?
- IT leadership and operations
- CISOs and information security officers
- Data protection officers
- Network and endpoint administrators
- Software asset management owners
What you will take away
- Recognise the forms shadow AI takes, from chat assistants and browser extensions to AI features inside SaaS tools.
- Use existing sources such as proxy logs, CASB, DNS and SaaS management to see which AI services are in use.
- Prioritise discovered usage by data risk.
- Apply an approach that offers secure alternatives rather than blanket blocking.
- Involve employees without creating a surveillance culture.
Agenda
Module 1
What shadow AI is
- Forms: chatbots, browser plugins, meeting bots, AI features in existing SaaS
- Why staff fall back on personal accounts
- Concrete risks: data leakage, training on inputs, prompt injection
Module 2
Creating visibility
- Analysing proxy, DNS and firewall logs
- Using CASB, DLP and SaaS management tools effectively
- Anonymous surveys as a complementary source
Module 3
Assessing and prioritising
- Risk matrix by data type and tool
- Exercise with real or anonymised findings
- Checking against policy and privacy requirements
Module 4
Govern rather than ban
- Secure alternatives: enterprise versions of ChatGPT, Claude, Microsoft 365 Copilot, Google Gemini or Mistral Le Chat
- A fast approval path for new tools
- Communication, training and works council involvement
Why bans alone do not work
Blocking AI tools simply moves usage to personal phones and laptops where IT has no visibility at all. People use AI because it saves them time.
A three-part approach works better: make usage visible, provide secure alternatives and communicate clear rules. Risk drops and the productivity gains stay inside the company.
Shadow AI and compliance
Uncontrolled AI use is not only a privacy issue. In an ISO 27001 ISMS, AI services belong in the asset inventory, and the EU AI Act expects deployers to ensure adequate AI literacy among staff. Both require knowing what is actually used.
- Leakage of confidential information
- Missing data processing agreements
- Unknown AI features in already approved software
- Gaps in asset and supplier management
Who runs it
The workshop is delivered by neooutput AI Trainers with a security background. jocapps, the company behind neooutput, is ISO 27001 and ISO 9001 certified.
Frequently asked questions
Do we need to buy new security tools?
Usually not. Many companies already have enough data in their proxy, firewall, endpoint management or Microsoft 365 tooling. We show how to analyse it in a targeted way.
Is log analysis subject to employee co-determination?
It can be, especially if conclusions about individuals are possible. We recommend aggregated analysis and early alignment with employee representatives and privacy.
What about AI features in tools we already use?
These are often overlooked. Part of the workshop covers how to spot and assess new AI features appearing in existing SaaS.
Workshop
Get in touch
Want to know which AI tools are really being used in your organisation?
Tell us a little about your team – we will get back to you within 24 hours with a proposal for content, format and dates.
Prefer to talk?
More workshops
- Workshops: EU AI Act & ComplianceAI Governance & Risk Management with ISO/IEC 42001: Workshop
- Workshops: EU AI Act & ComplianceCopyright & AI-Generated Content: Workshop for Companies
- Workshops: EU AI Act & ComplianceDo You Need to Label AI Content? Transparency Workshop
- Workshops: AI Tools at WorkLocal LLMs & Data Sovereignty: Ollama, Mistral, Llama Workshop
