AI literacy and certification in ICT and media.
ICT and media is the industry in which AI has both penetrated furthest and generated the most public debate. CBS reported 58.0% AI adoption in 2024 in the information and communication sector — the highest figure of any Dutch sector. At the same time, the AP warns of growing risks around deepfakes and synthetic content.
What is at play
What is at play in ICT and media.
In ICT companies, coding assistants are standard, with GitHub Copilot, Cursor and Microsoft Copilot the most widely used. NPO is developing its own recommendation algorithms; RTL uses AI for automatic thumbnail selection. ANP, NU.nl, Volkskrant, DPG Media and Mediahuis have all published formal AI guidelines. GPT-NL (TNO, NFI, SURF, supported by EZ with €13.5 million) is building a Dutch-language foundation model; NDP Nieuwsmedia contributed the archives of more than 30 titles.
Why AI literacy matters especially here
Transparency (obligation), DSA and copyright converge.
In this sector, four legal threads converge. First, the AI Act, under which biometrics, emotion recognition, critical digital infrastructure and recommendation systems on large platforms constitute high-risk or prohibited categories (Article 5 and Annex III). Second, Article 50 of the AI Act: providers and deployers must label synthetic output and make deepfakes identifiable, with an exception for text that has been editorially checked by a human. Third, the Digital Services Act and the Digital Markets Act. Fourth, copyright and the TDM exceptions under Articles 3–4 of the DSM Directive 2019/790.
The four SAIG levels for ICT and media
Four levels, one framework.
For support staff who use AI tools incidentally.
For junior developers, content creators, social media managers and editors who use AI in daily production.
Core level for software/data engineers, journalists, art directors and performance marketers who must independently account for AI output.
For ML engineers with model responsibility, CTOs, Chief AI Officers, editors-in-chief, DPOs and compliance officers (DSA, copyright, AI Act).
Which level fits which role?
Role and recommended level.
| Role | AI contact | Recommended level |
|---|---|---|
| Office assistant, support staff | incidental | Awareness Badge |
| Junior developer, content creator, social media manager | daily AI in production | Basis |
| Software/data engineer, journalist, art director | independently accountable for AI output | Practitioner |
| ML engineer, prompt engineer, UX designer | builds or steers AI systems | Practitioner |
| CTO, Chief AI Officer, editor-in-chief | governance, policy, copyright | Advanced |
| Compliance, DPO, FG | DSA, AI Act, GDPR | Advanced |
For organisations and for professionals
One standard, two tracks.
For ICT companies, broadcasters, publishers and agencies, SAIG offers a structured route to fulfilling Article 4 that aligns with the Article 50 transparency obligations and with the DSA requirements for very large platforms. For individual professionals, the certificate is a portable, independent proof of tested competence — relevant in a labour market where AI-related job titles are shifting at breakneck speed.
Frequently asked questions
Frequently asked questions about AI in ICT and media.
Does AI-generated content always have to be labelled?
From 2 August 2026, Article 50 applies. For text that has been editorially checked by a human, paragraph 4 contains an exception; for deepfakes and synthetic audio/video, labelling remains required.
Is SAIG mandatory for developers?
No. Article 4 requires a sufficient level of AI literacy; SAIG certification is one route.
Which level suits an ML engineer?
Practitioner; Advanced when bearing final responsibility for production models.
How does SAIG relate to the DSA?
Complementary. The DSA sets requirements for platforms; SAIG certifies the individuals who substantively shape compliance.
May an editorial team use ChatGPT?
Yes, provided it follows editorial guidelines, with human oversight and — for synthetic images or voices — with a transparency label.
What applies to copyright and training data?
The TDM exceptions from Articles 3–4 of the DSM Directive 2019/790 with opt-out for rights holders are leading. SAIG-Advanced tests this competence.
Next step
Schedule an orientation call.
We discuss what Article 4 specifically means for your organisation in ICT and media and which SAIG route fits your roles and risk profile.
Sources: CBS AI-monitor 2024; AP-rapportage AI & algoritmische risico's 2024; TNO over GPT-NL (juli 2025); NDP Nieuwsmedia; Villamedia overzicht AI-richtlijnen redacties; EUR-Lex Verordening (EU) 2024/1689; DSA Verordening (EU) 2022/2065.