Gesprächsreihe UNESCO Internet Universalitäts-Indikatoren - Gespräch zum Indikatorenbereich Zugang

Summary of the discussion (UNESCO Internet Universality Indicators – “Access” area)

1) Germany’s internet infrastructure: progress, but “fiber to the home” remains a core gap

• The video references a 2021 UNESCO Internet Universality Indicators report saying Germany was “lagging significantly” in broadband/fiber deployment because plans had been neglected for a long time.
• The discussion asks whether this is still true for 2026.

Philip Emann argues the picture is mixed:

• Positive development: Germany’s fiber expansion has gained momentum, with improving framework conditions (a new federal digital-infrastructure focus following ministry restructuring).
• Persistent gap: “fiber” can mean different things, and measurement/definitions matter:
  • Many homes may have fiber available along the street, but fewer have fiber connections inside homes/buildings.
  • Estimates discussed include:
    • ~50% of homes by 2025 with fiber-related build-out momentum
    • but only ~20–30% (likely ~25–30%) with direct “fiber to the home” connectivity.
• Adoption challenge: even with infrastructure, some people do not use it. This can relate to:
  • how technology is defined (e.g., “true” multi-core fiber vs hybrid copper/fiber setups),
  • contract and incentive structures.

Policy relevance:

• The TKG amendment is highlighted as a key context for getting fiber into buildings and reshaping access obligations.

---

2) Mobile networks, satellites, and “technology-neutral” access rights

The discussion expands beyond fiber:

• Mobile coverage is described as relatively good overall, with ongoing reforms aimed at helping operators expand networks and use frequencies more effectively.
• Satellite internet is increasingly important, particularly for:
  • industrial/B2B needs,
  • cross-border connectivity (e.g., connected vehicles switching networks while traveling).

Legal concept emphasized (Philip Emann):

• Germany’s Telecommunications Act (TKG) frames minimum service entitlement as technology-neutral access to bandwidth/latency, not a right to dig and install specific infrastructure.

Broader systems view (both contributors):

• Digital infrastructures are increasingly interconnected across cable, mobile, and satellite.
• This raises questions about digital sovereignty and dependence on providers.

---

3) AI is changing “access” and usage patterns—mostly via cloud/data centers rather than local devices

Valentina Kerst notes that although AI tools are widely used (the “ChatGPT moment”), access is uneven, especially across user groups:

• Premium users and corporate/scientific communities can use AI more intensively than the general public.

Infrastructure linkage (Kerst + team):

• They distinguish:
  • everyday AI use on phones (generally feasible over mobile networks),
  • heavier AI workloads (training/foundational models) that require major compute resources in data centers.

Key point:

• For widespread AI deployment, local execution on mobile devices is expected to remain a minority case.
• The central dependency shifts to where AI models run—i.e., data centers and compute infrastructure.

---

4) European/German “digital sovereignty” in AI: progress, but adoption and demand are still uncertain

Kerst responds to concerns about dependence on major AI providers outside Europe (especially the US and China):

• Europe can still respond through momentum in infrastructure, including:
  • mention of ~60 AI data centers under construction in Germany,
  • readiness of the Jülich supercomputer.
• The major uncertainty is whether capacity will be actually used.

Barriers discussed:

• The “login effect” / habituation: companies often stay with familiar US/Chinese ecosystems.

Reframing sovereignty:

• Digital sovereignty is not treated as 100% isolation, but as:
  • having choice,
  • reduced dependency,
  • enough European providers,
  • and willingness/ability to switch.

Emann’s cloud-market perspective:

• Short innovation cycles make lock-in less absolute.
• Many firms move toward multi-cloud strategies:
  • sovereign clouds for IP-sensitive tasks,
  • other systems for less sensitive needs.
• Specialized European/national providers may grow by succeeding in niches, then scaling later.

---

5) Equal access and skills: education gaps will widen as AI accelerates advantages

The conversation shifts from connectivity to knowledge, utility, and skills, using UNESCO-style evidence:

• Internet access correlates strongly with education level (higher for those with higher educational attainment than for those with secondary-level credentials).
• Employment vs unemployment shows lower internet use among unemployed people.

Justice/inclusion framing (Kerst):

• People with AI skills gain more, which may widen gaps.

Nuance with statistics (Emann):

• The Federal Statistical Office reports ~3% offline for ages 16–74.
• But the measure:
  • misses younger people (ages 8–16),
  • is less certain for older groups.

What “access” means in practice:

• It’s not just “being online,” but being able to use services effectively, for example:
  • why some people don’t adopt tools like digital identity wallets or online administrative services (they may prefer familiar in-person processes),
  • the need for media and IT competence to use AI safely and critically, including understanding limits such as hallucinations.

---

6) What digital/IT/media competence should look like in education (and why tech alone isn’t enough)

Both guests argue education needs to account for different target groups and life stages:

• Not “one-size-fits-all” training.
• Training should include pathways for:
  • basic learning,
  • further training,
  • advanced competence.

Kerst’s suggestions:

• Make skill-building attractive, using incentives/certificates (she notes she’s not necessarily in favor of overly rigid obligation-based approaches).
• Avoid simplistic metrics like only distributing devices.

Emann’s emphasis (media literacy):

• Focus on AI/digital competence in schools through:
  • source evaluation,
  • authorship and accountability,
  • cross-checking results,
  • cultural techniques for navigating information.

Reflection on past digitization (Corona period):

• Digital rollout sometimes replicated traditional frontal teaching rather than enabling new learning models (e.g., meaningful self-study resources).
• The pandemic is treated as a moment that made change possible, but it also showed the risk of “switching on/off” without thoughtful integration.

Overall direction (Kerst):

• Technology should be framed as a tool to use time more effectively, not as an attention-consuming “hamster wheel.”
• The education system and public discourse should re-align around structured, purposeful adoption in the AI era.

---

Presenters / contributors

• Valentina Kerst
• Philip Emann