6G Topic 04 — AI/ML-Native Air Interface

Putting learning inside the radio

In 5G, AI lives mostly in the network's brain — scheduling, optimisation, operations. The 6G idea is more radical: build machine learning into the air interface itself, so the link between phone and tower learns. The honest status is that this is a direction with real, completed study work behind specific functions — not a finished, AI-first radio.

AI/ML for Air Interface IMT-2030 — AI & Communication
Status. "AI-native" is an aspiration and an active study area, not a delivered design. Concrete progress exists for specific functions (CSI, beam management, positioning); the degree to which 6G is fundamentally AI-first is still under study.

Why the radio is a good place for ML

Classical radio algorithms are hand-designed from models — a channel estimator assumes a certain channel, a codebook assumes certain antenna geometry. Reality is messier than the models. Machine learning offers algorithms that adapt to the actual deployment from data. Three jobs are especially ripe, because they are high-volume, pattern-rich, and expensive to do optimally by hand: channel state feedback, beam management, and positioning.

requirement"AI & Communication" is one of the six IMT-2030 scenarios

ITU-R M.2160 names Artificial Intelligence and Communication as a headline 6G usage scenario — the network as a distributed compute/AI fabric. That makes AI a requirement-level driver for 6G, not an optional extra.

itu.int — IMT-2030 2023-11
candidate3GPP has studied ML for CSI, beams and positioning

The 6G AI/ML-for-air-interface direction builds directly on completed and ongoing 3GPP work (the Rel-18 study captured in TR 38.843, continued in Rel-19) covering CSI feedback/compression, beam management and positioning. These are the proving grounds; how much carries into normative 6G is a Release 21 question.

wirelessbrew.com 2026-03 secondary

The hard part: two-sided models

Channel-state-information (CSI) feedback is the flagship example. Today a phone measures the channel and reports a compressed summary so the tower can beamform. Replace the compressor with a neural encoder on the phone and a neural decoder at the tower and you can shrink the report dramatically. But now two different vendors' models — one in the phone, one in the base station — must agree, like two halves of a codec built by strangers. That "two-sided model" coordination (training, versioning, interoperability) is the genuinely novel standardisation problem 6G must solve.

Interactive — CSI feedback as a two-sided model illustrative concept — not a 6G spec value
64 of 512 raw coefficients
Illustrative only. The 512→latent numbers are a teaching example of the encode→feedback→decode idea, not a standardised 6G configuration. The real, hard, unsolved part is making the phone's encoder and the tower's decoder interoperate across vendors.
foundationCSI feedback and beam management already exist in 5G

The functions ML aims to improve are not new — 5G NR already has CSI reporting and a beam-management framework (SSB/CSI-RS beam sweeping, reporting, indication). Understanding the 5G versions is the fastest way to see precisely what ML changes and what stays the same.

ranbits — Foundations (5G beam sweep) 2025
AI/ML use cases for the air interface — 5G approach vs 6G candidate TR 38.843 (Rel-18) · wirelessbrew.com 2026-03 (secondary)
Use case 5G NR approach (foundation) AI/ML candidate approach Key challenge 3GPP study Status
CSI feedback compression Codebook-based Type I / Type II reporting; fixed compression Neural encoder (UE) + decoder (gNB) — data-driven; latent as small as 16 values Two-sided model: vendor interoperability, versioning, retraining TR 38.843 (Rel-18) — feasibility confirmed candidate
Beam management SSB / CSI-RS sweep → UE reports RSRP → gNB selects TCI state ML predicts best beam from history / position, reducing sweep overhead Generalisation across environments; beam-failure latency TR 38.843 (Rel-18) — studied candidate
Positioning OTDOA / UL-TDOA (Rel-16); DL-AoD / UL-AoA (Rel-17) ML exploits multipath fingerprinting for sub-meter NLOS accuracy Training data per deployment; NLOS vs LOS classification Rel-18/19 study candidate
Channel estimation Pilot-based MMSE / LS; interpolation across REs Model-aided / deep learning CE — better in sparse-pilot scenarios Complexity vs gain trade-off; low-latency inference on UE 6G study item candidate
Tracker — what 3GPP / ITU-R is doing here full tracker ↗
AI/ML for Air Interface (6G) candidate
AI/ML-native air interface (CSI, beam management, two-sided models)
Builds on the Rel-18 study (TR 38.843) and Rel-19 work on AI/ML for the air interface — CSI feedback/compression, beam management, positioning, and two-sided models — toward an AI-native 6G design. The degree of AI-native design in 6G is still under study.
wirelessbrew.com 2026-03 secondary
Study direction carried from Rel-18/19 AI/MLno % published

Where this connects

AI rides on the physical layer (it learns the waveform's channel) and is pulled by the requirements (AI & Communication is a named scenario). It also overlaps sensing, where ML interprets the echoes the radio collects.

Foundations Beam management — the most ML-ready 6G function — is concrete 5G NR today. See how SSB beam sweeping and TCI work before layering ML on top. 5G beam sweep →