Frontier technology governance
Frontier technologies such as advanced AI, engineered biology, quantum computing and autonomous robotics develop quickly, cross borders easily and combine in ways no single regulator oversees. Their governance mixes binding law, standards, export controls, voluntary commitments and international dialogue, often under deep uncertainty about capability and risk. This page sets out how governance gaps open at the frontier, how capabilities and controls are tracked, and what the bodies that hold the decisions need.
Regulatory sandbox board
Frontier technology governance · Regulatory sandbox · Cohort 3
Cohort
| Parametric micro-cover | Insurance | Agreed | 3 of 12 | Testing |
| Flood sensor data market | Data | Partial | 3 of 12 | Conditions |
| AI permit triage | Public service | Agreed | 7 of 12 | Testing |
| Drone deliveries, medical | Logistics | Agreed | 10 of 12 | Exit test |
Exit criteria
- Consumer harm
- none recorded
- Evidence
- independent evaluation
- Rule change
- regulator decides
- Wind-down
- plan on file
Sandbox cycle
- Applyeligibility
- Admitsafeguards
- Testbounded
- Evaluateindependent
- Exitor rule change
- Publishlessons
Fig.Regulatory sandbox board. Design wireframe with placeholder names and illustrative values, not a record or a live system.
The short answer.
Frontier technology governance is the mix of binding law, standards, export controls, voluntary commitments and international dialogue applied to fast-moving technologies such as advanced AI, engineered biology, quantum computing and autonomous robotics. These technologies cross borders easily and combine in ways no single regulator oversees.
People and hazards in the Mekong Basin.
Where people live, night-time lights and the floods and fires public institutions answer to, framed on the Mekong Basin as an example geography for this subject; each layer is a dated NASA product.
The interactive map needs JavaScript; the layer list and sources are below.
Layers 2
People and hazards
Keyboard, with the map focused
- Arrows
- Move
- Shift and arrows
- Turn and tilt (3D)
- + and −
- Zoom
- 0
- Reset the view
- N and P
- Next and previous station
- Enter
- Pin the station
- I
- Inspect a point
- M
- Measure
- C
- Compare dates
- Escape
- Close
Hold Ctrl or ⌘ and scroll to zoom
Latest available
NASA GIBS / EOSDIS · ASTER GDEM (METI, NASA) · Natural Earth
Loading the latest NASA imagery.
Drag to move. Ctrl or ⌘ and scroll, pinch or the buttons to zoom; hover for the values under the pointer; reset returns to the region.
Where people live, night-time lights and the floods and fires public institutions answer to, framed on the Mekong Basin.Satellite and model products as NASA publishes them, each with its own date; indicators, not assessments of any place; claim ceiling INTEGRATED.
Data: NASA GIBS · ASTER GDEM · Natural EarthSources and licences
- We acknowledge the use of imagery provided by services from NASA's Global Imagery Browse Services (GIBS), part of NASA's Earth Science Data and Information System (ESDIS). Colours redrawn from the published colour map; values averaged into H3 cells where shown (modified). NASA open data
- We acknowledge the use of imagery provided by services from NASA's Global Imagery Browse Services (GIBS), part of NASA's Earth Science Data and Information System (ESDIS). Colours redrawn from the published colour map; values averaged into H3 cells where shown (modified). NASA open data
- Center for International Earth Science Information Network (CIESIN), Columbia University. 2018. Gridded Population of the World, Version 4 (GPWv4): Population Density, Revision 11. NASA Socioeconomic Data and Applications Center (SEDAC). Licensed under CC BY 4.0. Colours redrawn from the published colour map; values averaged into H3 cells where shown (modified). CC BY 4.0
- VIIRS Black Marble night-time lights, NASA Goddard Space Flight Center; distributed by the NASA Level-1 and Atmosphere Archive and Distribution System (LAADS DAAC); imagery via NASA GIBS. Colours redrawn from the published colour map; values averaged into H3 cells where shown (modified). NASA open data
- VIIRS near-real-time products (flood, land surface temperature, snow cover, corrected reflectance), NASA LANCE; imagery via NASA GIBS. Colours redrawn from the published colour map; values averaged into H3 cells where shown (modified). NASA open data
- NASA Fire Information for Resource Management System (FIRMS), VIIRS 375 m active fire product, NASA LANCE; vector tiles via NASA GIBS. Colours redrawn from the published colour map; values averaged into H3 cells where shown (modified). NASA open data
- ASTER GDEM is a product of METI and NASA (ASTER Global Digital Elevation Model V003, NASA LP DAAC); imagery via NASA GIBS. Colours redrawn from the published colour map; values averaged into H3 cells where shown (modified). NASA open data
- Made with Natural Earth. Free vector and raster map data @ naturalearthdata.com. Public domain
Layers and dates
- Population density, 2020: Population: CIESIN, Columbia University, Gridded Population of the World v4.11 (2018), NASA SEDAC, CC BY 4.0 · static
- Night-time lights, daily: VIIRS Black Marble (VNP46A2), NASA LAADS DAAC · 12 Oct 2026
- Flood water, 3-day composite: VIIRS Global Flood Product, NASA LANCE · 14 Oct 2026
- Active fires and thermal anomalies, last day: VIIRS NOAA-20 375 m (FIRMS), NASA LANCE FIRMS · 13 Oct 2026
- True colour: VIIRS NOAA-20 corrected reflectance, NASA LANCE · 13 Oct 2026
Frontier technology governance: how the risk develops.
Crossover with AI, biotech, quantum and robotics. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.
Pace outruns process
Capabilities can change within months while legislation, standards and treaties take years. Rules written for one generation of a technology meet the next, and guidance fills the gap without binding force.
Convergence
AI accelerates biological design, quantum computing threatens current cryptography, and robotics gives software physical reach. Risks emerge at intersections that sit between the mandates of separate regulators.
Dual use and diffusion
The same models, tools and know-how serve beneficial and harmful ends. Openness speeds research and scrutiny and also lowers barriers to misuse, while export controls reach only part of the chain.
Asymmetric knowledge
Developers know more about capabilities and incidents than governments do. Without access, evaluation capacity and reporting duties, authorities govern from public claims and media reports.
Terms used on this page.
Each term as this page uses it. Where a standard-setting body defines the term, the wording follows that body.
| Term | Definition |
|---|---|
| Training compute | The total computation used to train an AI model, used as a rough proxy for its scale. |
| Frontier safety framework | A developer's published policy on evaluating and mitigating severe risks from its most capable models. |
| Wassenaar Arrangement | A multilateral export control regime for conventional arms and dual-use goods and technologies. |
What is measured, and how well.
Each indicator carries its unit, its cadence, the kind of source behind it and the evidence label a value would carry. No values are shown: a programme supplies them under its own authority.
| Indicator | Unit | Cadence | Source type | Label |
|---|---|---|---|---|
| Training compute of notable AI models | floating-point operations | per model release | Developer disclosures and research trackers | ESTIMATED |
| Jurisdictions with binding AI legislation in force | count of jurisdictions | annual | Official gazettes and legislative trackers | OFFICIAL |
| Published international standards on AI, quantum and robotics | count by committee | quarterly | ISO and IEC catalogues | OFFICIAL |
| Frontier developers publishing safety frameworks | count of developers | annual | Developer publications | DECLARED |
| Export control entries for emerging technologies | count of controlled items | annual | Export control regimes and national lists | OFFICIAL |
| Reported AI incidents and hazards | count by harm type | monthly | Incident monitors | REPORTED |
| Gene synthesis providers applying sequence screening | count of providers | annual | Industry consortium membership records | DECLARED |
From sources to decisions.
The topic as a computational graph: named sources feed the methods, an evidence review gates the outputs, and the decisions stay with the authorities that hold them.
Fig. 02Nodes carry their type, two metric rows and a state; connectors show what feeds what; dashed frames name the owners; the timeline is the cadence. Design of the pipeline, not a record of a run.
Where the evidence comes from.
Named public datasets and institutional records, each with its licence class. Sources by agreement are used only under their terms and never republished.
| Source | What it gives | Licence class |
|---|---|---|
| Stanford AI Index Report | Tracking of AI research, models, investment, policy and public opinion | Open, CC BY-ND |
| International AI Safety Report | Independent scientific synthesis on general-purpose AI capabilities and risks | Free to read, terms per publication |
| OECD.AI Policy Observatory and AI Incidents Monitor | National AI policies, compute indicators and incident data | Free, attribution (OECD terms) |
| ISO and IEC catalogues (JTC 1/SC 42 and related committees) | Published and in-progress standards on AI, quantum and robotics | Catalogue free to search; standard texts under paid licence |
| Wassenaar Arrangement control lists | Dual-use goods and technologies subject to export control by participating states | Public document |
| Epoch AI data on notable models | Records of notable AI models, training compute and hardware | Open, CC BY |
| International Gene Synthesis Consortium | Harmonised screening protocol for synthetic DNA orders | Public document |
Who decides, and what the evidence serves.
Innovation Nexus prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.
| Who decides | The question | Evidence this page serves |
|---|---|---|
| European Commission AI Office and national AI authorities | How to apply obligations for general-purpose AI models with systemic risk | Capability trackers; comparative safety frameworks; codes of practice |
| National AI safety and security institutes | Which capabilities to evaluate and how to share results with regulators | Comparisons of evaluation methods; capability and incident tracking |
| Export control authorities of Wassenaar participating states | Which emerging technologies to add to control lists, and at what thresholds | Control list comparisons; diffusion and dual-use analyses |
| Standards bodies (ISO/IEC JTC 1, IEEE, CEN-CENELEC) | Where consensus standards can support law and where research is too unsettled to standardise | Standards gap maps against legal requirements |
| Health and biosecurity authorities | How to oversee AI-enabled biological design and DNA synthesis | Convergence scenarios; comparisons of screening practice |
Outputs this topic produces
- Governance maps showing which body covers which frontier risk, and where none does
- Capability and control trackers with every entry labelled by source
- Convergence scenarios for AI with biology, quantum computing and robotics
- Dialogue briefs that let regulators, standards bodies and developers compare frameworks
What experts ask.
Frontier questions concern speed, convergence and fragmented authority.
- How do advances in one technology accelerate others, for example AI in biological design?
- Which harms emerge from convergence that no single sectoral regulator covers?
- How can governance keep pace without freezing useful development?
- Which international mechanisms exist, and where are the gaps?
- How are dual-use risks balanced against openness in research?
Mechanisms and drivers.
Frontier technologies are general purpose: they can be applied across many sectors, which makes them hard to govern through sector-by-sector rules. Convergence multiplies both benefit and risk; AI accelerates materials discovery and biological design, robotics brings AI into physical space, and quantum computing may eventually affect cryptography and simulation.
Governance is spread across regulators, standards bodies, funders, export control authorities and international forums, each with partial mandates. Firms often hold the most technical knowledge. Gaps arise where risks fall between mandates or cross borders faster than authorities can coordinate, and information asymmetry between developers and overseers makes those gaps harder to see.
How it is studied, and where the methods stop.
| Method or data | What it shows | Limits |
|---|---|---|
| Horizon scanning and foresight | Emerging developments and convergences | Foresight is not prediction |
| Governance mapping | Which institutions cover which risks | Formal mandates may differ from practice |
| Capability evaluation | What frontier systems can actually do | Evaluations lag rapid development |
| Comparative policy analysis | How jurisdictions approach frontier technology | Approaches change quickly |
How it is studied, and what the methods cannot say.
- Artificial intelligence and machine learningCoding policy and legal texts · Mapping patents and papers to technologiesApplied here
- Policy, institutional and decision scienceMandate and gap mapping · Programme evaluation design · Comparative regulatory analysisApplied here
- Systems engineeringRequirements traceability for AI governance · Readiness gates for transfer programmesApplied here
Questions and answers.
Why use training compute as an indicator?
Who governs frontier technologies internationally?
Does Innovation Nexus evaluate AI models?
Why are frontier technologies hard to regulate?
Is international governance of frontier technology possible?
Does Innovation Nexus evaluate frontier systems?
Around Innovation Nexus.
Bring your expertise to this field.
The current service level is R0: information and enquiries. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.