Europe, Japan and South Korea Explore a 3GPP-Style AI Standards Partnership: What the AIPP Letter of Intent Means
Key takeaway: On 2 November 2026, Europe’s ETSI and the ICT standards bodies of Japan (TTC) and South Korea (TTA) signed a Letter of Intent to explore a possible AI Partnership Project (AIPP), modelled on the 3GPP mobile-standards partnership. No AIPP standards exist yet: this is an agreement to explore, not a launch.
On 2 November 2026, the European Telecommunications Standards Institute (ETSI), Japan’s Telecommunication Technology Committee (TTC) and South Korea’s Telecommunications Technology Association (TTA) signed a Letter of Intent to explore a possible AI Partnership Project, or AIPP. ETSI’s announcement, from Sophia Antipolis in France, says the project would draw on the collaborative model of 3GPP, the partnership behind successive generations of mobile standards.
What is confirmed — and what is not
Confirmed: The three bodies agreed to explore a possible AIPP, and the Letter of Intent was signed on 2 November 2026. ETSI describes it as industry-led, giving industry, academia and research a platform to align priorities and speed up standards development. ETSI is one of only three bodies recognised by the EU as a European Standards Organisation, with over 900 members from 60 countries.
Not confirmed: No start date, working groups, member companies or first standards have been named, and no other region has joined. ETSI’s director general invited “partners and standards shapers worldwide” to join, so membership is unsettled. Any claim that AIPP has already published AI standards is wrong; the next milestone is a published terms of reference or work programme.
Why the 3GPP model is the template
3GPP, the 3rd Generation Partnership Project, was founded in December 1998 by national and regional standards organisations to build a globally applicable third-generation mobile system. It later produced the specifications behind 4G LTE and 5G New Radio, used by billions. Its structure is the point: regional bodies act as organisational partners, set policy and scope, and turn partnership specifications into regional standards, while companies, operators and researchers contribute proposals through working groups.
TTC and TTA are already 3GPP organisational partners alongside ETSI, so the three bodies already cooperate this way. The difference is difficulty: mobile standards describe how radios interoperate, while AI standards must cover data quality, model behaviour, safety testing and governance.
| 3GPP (mobile) | Proposed AIPP (AI) | |
|---|---|---|
| Status | Operating since December 1998 | Exploration only; Letter of Intent signed 2 November 2026 |
| Partners | Seven regional standards bodies, including ETSI, TTC and TTA | ETSI, TTC and TTA so far, with an open invitation to others |
| Output | Specifications behind 3G, 4G LTE and 5G | No standards yet; scope still to be defined |
| Core challenge | Interoperability of radios and networks | Trust, safety, data documentation and evaluation of AI systems |
What standards could AIPP eventually cover?
ETSI’s announcement lists no technical scope, so any topic list is speculation. At an ETSI workshop in May 2026, Korean contributors pointed to gaps for “physical AI” — robots and systems acting in the real world — including validity criteria for synthetic data, documentation of data and model production, and evaluation metrics for world models. Those were proposals, not agreed AIPP work items.
ETSI already works in this area. On 29 October 2026 it published a specification on combating deepfakes, ETSI TS 104 275, which takes a preventive approach — safeguards before release, human review, content labelling and incident investigation — rather than relying only on detecting manipulated content afterwards.
How does AIPP compare with existing AI standards efforts?
AIPP would not arrive in an empty field: ISO and IEC already publish AI standards on terminology, data quality and AI management systems, and the EU AI Act creates requirements standards can help companies meet. ETSI positions AIPP as complementing that ecosystem. Its possible advantage is speed — industry-led partnerships like 3GPP move faster because builders sit at the table. Its risk is the opposite: standards a few regions or large vendors dominate may simply be ignored.
What would it cost, and what must happen first?
No budget for AIPP has been announced, so no figure can be quoted. In the 3GPP model, partner bodies fund coordination while companies bear their own costs of sending engineers and implementing specifications. For most companies the real cost is implementation — testing, documenting data and model behaviour, and re-certifying when a standard is revised — and smaller developers would feel that most.
Three prerequisites come before any specification: agreement on scope; agreement on working procedures, including intellectual-property rules for contributed technology; and participants beyond the three founders willing to contribute and test implementations. Until those are visible, AIPP remains a letter, not a programme.
Risks, limits and what to watch
The largest risk is fragmentation by another name: if Europe, Japan and South Korea align on one set of AI standards while the United States, China and others pursue different ones, developers could face competing compliance regimes instead of one rulebook. A second risk is pace — AI changes in months, standards traditionally take years. A third limit is enforcement: standards are voluntary unless regulators or large buyers require them.
Analysis: The significance here is not a standard — there is none yet — but a signal. The October 2026 agreement by Google, Microsoft, OpenAI and Anthropic on a shared robots.txt signal for AI training showed even rivals want common mechanisms over proprietary patchworks. AIPP is the standards-body equivalent. If it widens and moves quickly, smaller builders get one target instead of three; if it stays a three-body exploration, it joins a long list of well-meaning talks. The next document — scope, or silence — will tell us which.
Watch for: AIPP terms of reference or a work programme; whether bodies from North America or elsewhere join or launch a rival effort; and whether regulators reference any future AIPP output.
For context, see our coverage of the shared robots.txt opt-out agreed by the major AI companies and our guide to why long-term impact analysis matters in AI ethics.
Sources
- ETSI — ETSI, TTC and TTA Explore New Global Partnership for Standardisation in Artificial Intelligence (2 November 2026)
- Telecompaper — ETSI to explore 3GPP-inspired AI partnership with Japanese, South Korean peers (2 November 2026)
- 3GPP Working Procedures — partnership structure and purpose
Frequently asked questions
What is the AI Partnership Project (AIPP)?
AIPP is a possible industry-led partnership for AI standardisation proposed by ETSI, Japan’s TTC and South Korea’s TTA. On 2 November 2026 the three bodies signed a Letter of Intent to explore creating it, modelled on the 3GPP mobile-standards partnership. It is not yet an operating organisation and has published no standards.
Is AIPP already a standards body?
No. The Letter of Intent only agrees to explore the options for a possible partnership. No start date, working groups, member companies or first standards have been announced. A future agreement on scope and working procedures would be needed before AIPP could begin standard-setting work.
How is AIPP different from 3GPP?
3GPP is an operating partnership founded in December 1998 whose specifications underpin 3G, 4G LTE and 5G mobile networks. AIPP would apply a similar partnership model to AI, but its subject is harder: data quality, model safety, documentation and evaluation, rather than radio and network interoperability. TTC and TTA are already 3GPP partners alongside ETSI.
What standards might AIPP eventually produce?
No scope has been agreed, so this is not settled. Workshop proposals discussed at ETSI in May 2026 pointed to gaps for physical AI, including validity criteria for synthetic data, documentation of data and models, and evaluation metrics for world models. Those were proposals, not agreed AIPP work items.
What does AIPP mean for ordinary users and developers?
In the short term, nothing changes: no new rules apply to any product. In the longer term, if AIPP produces widely adopted standards and regulators or large buyers require them, developers could gain one common set of requirements for Europe, Japan and South Korea instead of separate national rules, while users could get AI systems tested against shared safety and documentation expectations.
