About Human-Level AI

A global intelligence publication mapping the rise of AGI.

AGI is emerging through a combination of technological breakthroughs, institutional decisions, economic shifts, geopolitical competition and societal change. We track these developments, examine their implications and make the changing global picture understandable.

Current public release20260516-161255

Observatory snapshot exported 16 Nov 2026, 5:13 a.m. ET.

Mission

The rise of AGI is not happening everywhere at once. Our job is to show what is changing, where it is happening and why it matters: for people, institutions, economies, security and the future of society. We aim to make consequential developments understandable without making them sensational.

How the publication works

The Observatory shows what is happening. The journalism explains why it matters. The human perspective reminds us who is affected. Private editorial workflows are kept separate from public delivery: the public site is rendered only from approved releases.

Methodology

What is published

Signals an editor has reviewed and published, and items from monitored official sources that passed an automated AI check, which are labelled “Published automatically”. Each one cites at least one public source, or is labelled as editorial analysis with its methodology. Withdrawn signals are removed.

Facts, interpretation and inference

Every signal carries one of three labels: source-reported fact, editorial synthesis or inference. Evidence is described qualitatively, for example “single source” or “multiple independent sources”, with a written rationale. We do not publish numeric confidence scores.

Locations

We use the least precise location that is still useful. Approximate markers (region or country) do not mark an exact site. Signals without an established location are listed but not pinned. Coordinates are never guessed and personal locations are never mapped.

Timestamps

Event time, source publication time, retrieval time and the time of our last source check are shown separately and are never treated as interchangeable. Missing values are shown as unknown.

Freshness

Source freshness is reported on its own. The site checks for new data periodically; a working update check does not make stale sources current, and an old snapshot is labelled stale rather than presented as current intelligence.

Risk index

An experimental risk index exists in the Risk Atlas. Its inputs, weighting, calculation frequency and limitations have not yet been published, so it is not used as a measure of global AI risk on the homepage or treated as authoritative anywhere on the site.

Editorial standards

Evidence first

Reporting and analysis are backed by cited sources with transparent provenance and licence credits.

Clear separation

Verified facts, reported claims and editorial interpretation are labelled, never blended.

No fabricated data

Counts, timestamps and locations on the site come from published artifacts only. Coverage limits stay visible.

Corrections

Corrections are recorded on the affected signal with their date, and withdrawn signals are removed from the map.

Coverage

Risk Atlas >

Eight monitoring domains. 81 signals: 79 reviewed by editors, 2 published automatically after an AI check. 10 sources are monitored for new items; each is checked before it is published. This is a curated selection, not a complete record of global events.

  • AI Capabilities & Risk

    Advances in AI capability, alignment, autonomy and potential loss of control.

    38 published signals
  • Biosecurity

    The intersection of advanced AI, biological research and emerging biosecurity concerns.

    1 published signal
  • Cyber Warfare

    AI-enabled cyber operations, digital vulnerabilities and geopolitical conflict.

    29 published signals
  • Cognitive Manipulation

    Information integrity, persuasion, synthetic media and the manipulation of public perception.

    4 published signals
  • Surveillance State

    AI-enabled monitoring, identification, control and the implications for civil liberties.

    4 published signals
  • Energy Systems

    The energy infrastructure and resources supporting increasingly powerful AI systems.

    1 published signal
  • Environment

    The environmental consequences of AI infrastructure, development and deployment.

    2 published signals
  • Transhumanism

    Human enhancement, human-machine integration and the changing boundaries of human capability.

    2 published signals

Monitored sources: arXiv preprints (cs.CY, cs.HC), ECDC news, European Data Protection Board, Electronic Frontier Foundation (updates), EFSA news, Epoch AI, U.S. FDA press announcements, NOAA news, UK AI Security Institute, UK National Cyber Security Centre. This is a curated selection, not a complete record of global events.

Collect

Monitored official sources and the legacy archive.

Check

Automated AI relevance and evidence check on every ingested item.

Review

Editorial review and classification before public release.

Publish

Static artifacts rendered to the public site from approved releases.