{"ontology":"https://layer.re/machine/schema.json","epistemic_model":"evidence-backed-temporal-layer-graph","canonical":true,"type":"layer_history","layer":{"id":"layer-human","slug":"human","name":"Human","description":"Human capability, labor, wellbeing and institutions.","machine_description":"Human capability, labor, wellbeing and institutions.","active":1},"snapshots":[{"id":"state-human-2022q1","layer_id":"layer-human","as_of":"2022-03-31 23:59:59.000","summary":"Humans were becoming more directly embedded in model behavior through preference labeling and feedback, while institutions were beginning to formalize AI risk to individuals and society.","dependencies":["layer-models","layer-governance","human labelers and users"],"constraints":["scalable supervision","bias and representativeness","uncertain social impacts"],"emerging_capabilities":["human-feedback post-training","structured societal-risk framing"],"underdeveloped_components":["scalable oversight for more capable systems"],"uncertainties":["human preference data may not represent broader populations or future use contexts"],"confidence":0.89,"previous_snapshot_id":null,"mapping_mode":"retrospective","reconstructed_at":"2026-10-01T17:34:32.000Z","created_at":"2026-10-01T18:05:36.397Z","updated_at":"2026-10-01T18:05:36.397Z"},{"id":"state-human-2022q2","layer_id":"layer-human","as_of":"2022-06-30 23:59:59.000","summary":"Humans became more visibly downstream users and workflow partners of generative models, not only supervisors or labelers upstream of training.","dependencies":["layer-models","layer-software","layer-governance"],"constraints":["trust","output correctness","misuse","workflow adaptation"],"emerging_capabilities":["creative text-to-image workflows","AI pair programming"],"underdeveloped_components":["clear labor-productivity evidence","stable social expectations for AI-generated work"],"uncertainties":["vendor adoption metrics may not represent durable use or net benefit"],"confidence":0.93,"previous_snapshot_id":"state-human-2022q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-01T18:14:00.000Z","created_at":"2026-10-01T19:22:17.766Z","updated_at":"2026-10-01T19:22:17.766Z"},{"id":"state-human-2022q3","layer_id":"layer-human","as_of":"2022-09-30 23:59:59.000","summary":"Humans interacted with AI across larger public generative surfaces and public dialogue-agent experiments while open speech, translation and image models widened the set of directly usable model capabilities.","dependencies":["layer-models","layer-software","layer-governance","layer-open-source"],"constraints":["trust","output correctness","misuse","accessibility","workflow adaptation"],"emerging_capabilities":["mass-access text-to-image generation","public feedback-driven chatbot research","open multilingual speech and translation tools"],"underdeveloped_components":["independent productivity evidence","stable social norms for generated media","reliable factual dialogue"],"uncertainties":["vendor-reported usage did not establish durable productivity or representative adoption"],"confidence":0.94,"previous_snapshot_id":"state-human-2022q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-02T15:12:29.000Z","created_at":"2026-10-02T16:08:16.259Z","updated_at":"2026-10-02T16:08:16.259Z"},{"id":"state-human-2022q4","layer_id":"layer-human","as_of":"2022-12-31 23:59:59.000","summary":"Humans gained a lower-friction general conversational interface to a capable instruction-following model while developers and organizations gained additional integration and administration surfaces for generative AI and coding assistance.","dependencies":["layer-models","layer-software","layer-governance"],"constraints":["trust","hallucination","harmful outputs","workflow adaptation","access and moderation"],"emerging_capabilities":["free multi-turn conversational research preview","developer-integrated image generation","organization-managed coding assistance"],"underdeveloped_components":["independent productivity evidence","reliable factual dialogue","stable norms for AI-generated work"],"uncertainties":["Q4 evidence does not justify importing later ChatGPT adoption, labor effects or institutional dependence"],"confidence":0.97,"previous_snapshot_id":"state-human-2022q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-03T14:26:11.000Z","created_at":"2026-10-03T14:45:10.928Z","updated_at":"2026-10-03T14:45:10.928Z"},{"id":"state-human-2023q1","layer_id":"layer-human","as_of":"2023-03-31 23:59:59.000","summary":"Humans encountered conversational generative AI across a wider set of products and providers, including search/browser, assistant interfaces and downstream applications, while access and reliability remained uneven.","dependencies":["layer-models","layer-software","layer-governance"],"constraints":["trust","hallucination","privacy","access gating","workflow adaptation"],"emerging_capabilities":["AI-assisted search/browser use","multiple assistant providers","GPT-4 text access","emerging enterprise-context assistants"],"underdeveloped_components":["independent productivity evidence","reliable factual dialogue","stable institutional norms","broad permission-safe enterprise deployment"],"uncertainties":["Q1 product availability does not establish later mass adoption or labor effects."],"confidence":0.97,"previous_snapshot_id":"state-human-2022q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-05T15:23:51.000Z","created_at":"2026-10-05T18:17:53.803Z","updated_at":"2026-10-05T18:17:53.803Z"},{"id":"state-human-2023q2","layer_id":"layer-human","as_of":"2023-06-30 23:59:59.000","summary":"Human access broadened geographically and functionally through wider assistant availability, long-document interaction, tool-enabled chat and visible data controls.","dependencies":["layer-models","layer-software","layer-governance"],"constraints":["trust","privacy","tool risk","hallucination"],"emerging_capabilities":["broader Bard access","long-document interaction","wider plugins","data controls"],"underdeveloped_components":["reliable factual interaction","safe delegated action"],"uncertainties":["availability differs from impact"],"confidence":0.96,"previous_snapshot_id":"state-human-2023q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.612Z","updated_at":"2026-10-07T14:04:48.612Z"},{"id":"state-human-2023q3","layer_id":"layer-human","as_of":"2023-09-30 23:59:59.000","summary":"Human interaction with AI became more embedded in ordinary work and computing through enterprise deployment, desktop OS integration, advanced data analysis, and the beginning of image and voice interaction.","dependencies":["layer-models","layer-software","layer-governance"],"constraints":["trust","privacy","organizational policy","multimodal errors","tool risk"],"emerging_capabilities":["enterprise assistant access","Advanced Data Analysis","Windows Copilot preview","voice/image interaction rollout"],"underdeveloped_components":["stable multimodal norms","safe delegated action","independent productivity evidence"],"uncertainties":["availability does not prove sustained organizational adoption or productivity effects"],"confidence":0.97,"previous_snapshot_id":"state-human-2023q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T16:04:51.000Z","created_at":"2026-10-07T16:56:16.870Z","updated_at":"2026-10-07T16:56:16.870Z"},{"id":"state-human-2023q4","layer_id":"layer-human","as_of":"2023-12-31 23:59:59.000","summary":"Human interaction with AI became more configurable and organizationally embedded through Microsoft 365 Copilot GA, custom GPTs and expanding multimodal assistants.","dependencies":["layer-models","layer-software","layer-governance"],"constraints":["trust","privacy","organizational policy","tool risk"],"emerging_capabilities":["enterprise Copilot GA","custom GPT creation","vision/audio APIs","Gemini-powered products"],"underdeveloped_components":["safe generalized delegation","independent productivity evidence"],"uncertainties":["Availability and configurability do not establish universal adoption"],"confidence":0.97,"previous_snapshot_id":"state-human-2023q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-08T15:50:00.000Z","created_at":"2026-10-08T17:00:00.134Z","updated_at":"2026-10-08T17:00:00.134Z"},{"id":"state-human-2024q1","layer_id":"layer-human","as_of":"2024-03-31 23:59:59.000","summary":"Users gained discoverable customized assistants, another paid frontier assistant and a limited cross-chat memory test; this shows available interaction choices rather than population-wide adoption.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["GPT Store starts rolling out","Gemini Advanced exposes Ultra 1.0 to subscribers","ChatGPT tests cross-chat memory with user controls","Claude 3 Opus and Sonnet become available with vision"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["Distribution of customized assistants does not establish autonomous execution or a deployed creator-payment system.","Access is plan and market bounded; it does not demonstrate widespread adoption.","Exclude September 2024 and 2025 updates; remembered preferences are not autonomous background work.","Haiku was announced for later availability; benchmark leadership and near-human intelligence remain vendor claims."],"confidence":0.93,"previous_snapshot_id":"state-human-2023q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-09T21:04:17.186Z","updated_at":"2026-10-09T21:04:17.186Z"},{"id":"state-human-2024q2","layer_id":"layer-human","as_of":"2024-06-30 23:59:59.000","summary":"Available user interfaces expanded toward free-tier multimodal assistance, local AI PCs and editable outputs; Recall's deferral preserved a concrete availability boundary.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["GPT-4o text and vision roll out","Recall broad preview is deferred","Copilot+ PCs become generally available","Claude introduces an artifact workspace beside chat"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["The announced native audio/video experience was not generally available at launch.","Copilot+ PC availability must not be used to imply Recall shipped with those PCs.","Hardware release is not population-wide use; Recall was deferred and not part of broad launch availability.","An editable artifact workspace does not establish autonomous execution; current source displays June 21 while same release is independently dated June 20 by AWS."],"confidence":0.93,"previous_snapshot_id":"state-human-2024q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-10T07:29:06.787Z","updated_at":"2026-10-10T07:29:06.787Z"},{"id":"state-human-2024q4","layer_id":"layer-human","as_of":"2024-12-31 23:59:59.000","summary":"Users gained editable collaboration, OS beta assistance, cited search and paid video tools; availability does not establish population-wide impact.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["Canvas introduces a collaborative writing and coding beta","Apple Intelligence begins eligible-device beta rollout","ChatGPT Search moves beyond SearchGPT prototype","Sora video product begins rollout"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["An editable workspace is not unsupervised software execution.","Future Siri cross-app actions and later language coverage are not launch capabilities.","Later free and logged-out access must not be read backward into October.","Generated video is not reliable physics simulation or a controllable game engine."],"confidence":0.93,"previous_snapshot_id":"state-human-2024q2","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-10T08:02:14.505Z","updated_at":"2026-10-10T08:02:14.505Z"},{"id":"state-human-2025q1","layer_id":"layer-human","as_of":"2025-03-31 23:59:59.000","summary":"Eligible users gained task-oriented browser, research and coding interfaces; errors and supervision remained material.","dependencies":[],"constraints":["availability scope","reliability and implementation limits"],"emerging_capabilities":["Operator begins US Pro research preview","Deep research launches for bounded web investigations","Claude 3.7 and Claude Code research preview launch"],"underdeveloped_components":["generalization beyond the selected evidence"],"uncertainties":["Login takeover, purchase confirmations and limited access are not generalized delegated authority or continuously autonomous operation.","Cited reports can contain errors; a research session is not general background autonomy. Later upgrades are excluded.","Model availability differs from coding-agent preview; autonomous coding claims do not establish unsupervised reliability."],"confidence":0.93,"previous_snapshot_id":"state-human-2024q4","mapping_mode":"retrospective","reconstructed_at":"2026-10-09T19:12:47.000Z","created_at":"2026-10-10T08:11:02.815Z","updated_at":"2026-10-10T08:11:02.815Z"},{"id":"state-human-2025q3","layer_id":"layer-human","as_of":"2025-09-30 23:59:59.000","summary":"Measured developer outcomes differ from perceived and advertised productivity. The controlled experienced-developer study supports a local slowdown, not a general conclusion about all AI-assisted work.","dependencies":[],"constraints":["The measured nineteen-percent slowdown is specific to this sample and setting, not every developer or later model."],"emerging_capabilities":["METR publishes experienced-developer productivity experiment"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-human-2025q1","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T08:31:22.792Z","updated_at":"2026-10-10T08:31:22.792Z"},{"id":"state-human-2026q1","layer_id":"layer-human","as_of":"2026-03-31 23:59:59.000","summary":"This period adds specific human evidence through Cowork enters desktop research preview; Codex desktop app launches on macOS. Maturity remains specific to the cited products and programs; this synthesis does not infer sector-wide adoption.","dependencies":[],"constraints":["Experimental desktop work is not general availability or proof of dependable unsupervised workplace autonomy.","The March Windows update is separate; long-running workflow claims do not quantify independent reliability."],"emerging_capabilities":["Cowork enters desktop research preview","Codex desktop app launches on macOS"],"underdeveloped_components":["Independent measurement of deployment scope and outcomes"],"uncertainties":["The referenced historical predecessor belongs to the reviewed chronology; the live history must be checked before ingestion."],"confidence":0.9,"previous_snapshot_id":"state-human-2025q3","mapping_mode":"retrospective","reconstructed_at":"2026-10-10T06:24:42.000Z","created_at":"2026-10-10T08:45:32.710Z","updated_at":"2026-10-10T08:45:32.710Z"},{"id":"state-human-20260930-initial","layer_id":"layer-human","as_of":"2026-09-30 12:45:00.000","summary":"AI is moving further into ordinary institutional and political debate, increasing pressure for understandable public choices around safety, work, infrastructure and accountability.","dependencies":["layer-ai","layer-governance","layer-energy"],"constraints":["uneven public understanding","rapid technical change","distribution of costs and benefits"],"emerging_capabilities":["more accessible AI work tools","public debate over AI rules"],"underdeveloped_components":["shared social expectations for autonomous systems"],"uncertainties":["labor effects","public trust","political durability of current policy positions"],"confidence":0.82,"previous_snapshot_id":null,"mapping_mode":"contemporaneous","reconstructed_at":null,"created_at":"2026-09-30T12:54:32.493Z","updated_at":"2026-09-30T12:54:32.493Z"},{"id":"state-human-20261001-human-context","layer_id":"layer-human","as_of":"2026-10-01 19:31:00.000","summary":"Human-specific context is becoming increasingly machine-addressable, persistent and reusable across AI interactions, allowing agents to reason over preferences, history, constraints and current circumstances without implying a separate canonical Human Context layer.","dependencies":["layer-ai","layer-agents","layer-identity","layer-permissions"],"constraints":["user control","context accuracy","privacy","portability"],"emerging_capabilities":["portable personal context","cross-session continuity","agent-readable human preferences and constraints"],"underdeveloped_components":["cross-provider interoperability","standardized provenance and correction semantics"],"uncertainties":["whether human context stabilizes as an independent layer or remains a composition of existing layers"],"confidence":0.88,"previous_snapshot_id":"state-human-20260930-initial","mapping_mode":"contemporaneous","reconstructed_at":null,"created_at":"2026-10-01T20:08:12.916Z","updated_at":"2026-10-01T20:08:12.916Z"},{"id":"state-human-20261004-deployment-capacity","layer_id":"layer-human","as_of":"2026-10-04 20:12:00.000","summary":"Human capability at the model-to-enterprise boundary is becoming an explicit investment target, while insider safety concerns highlight organizational capacity as part of the challenge of deploying increasingly capable systems.","dependencies":["layer-models","layer-agents","layer-software","layer-governance"],"constraints":["scarcity of experienced deployment engineers","organizational process scaling","training and validation lag"],"emerging_capabilities":["specialized frontier deployment-engineering workforce"],"underdeveloped_components":["independent outcome evidence for deployment-training programs","validated organizational safety-process scaling methods"],"uncertainties":["whether FDE scarcity is a dominant bottleneck across sectors","how broadly Robinson's concerns generalize"],"confidence":0.88,"previous_snapshot_id":null,"mapping_mode":"contemporaneous","reconstructed_at":null,"created_at":"2026-10-04T21:57:54.258Z","updated_at":"2026-10-04T21:57:54.258Z"},{"id":"state-human-20261009-verification-throughput","layer_id":"layer-human","as_of":"2026-10-09 04:25:30.000","summary":"Human verification capacity is becoming a measurable bottleneck in AI-accelerated software security. Anthropic's open-source pipeline generated 29,439 candidate findings while only 6,123 had reached external human triage by October 2; 5,674 of those were confirmed valid. Independent maintainer release records also show real fixes from the program. The evidence is strong for a verification-throughput constraint, but it is concentrated in one vendor-led cybersecurity program and does not establish the same bottleneck across all AI-assisted knowledge work.","dependencies":["layer-models","layer-software","layer-open-source","layer-research"],"constraints":["expert triage capacity","maintainer attention","safe patch validation","disclosure coordination","severity calibration"],"emerging_capabilities":["AI-generated vulnerability reports with reproducers and candidate patches","high-throughput machine-assisted security discovery"],"underdeveloped_components":["automated trustworthy triage","safe automated remediation","maintainer throughput at discovery scale","independent cross-provider replication"],"uncertainties":["how representative Anthropic's pipeline is of other domains and providers","how raw unreviewed scanner precision changes outside selected high-severity samples","how quickly disclosed findings convert into deployed fixes"],"confidence":0.95,"previous_snapshot_id":"state-human-20261004-deployment-capacity","mapping_mode":"contemporaneous","reconstructed_at":null,"created_at":"2026-10-09T04:28:55.055Z","updated_at":"2026-10-09T04:28:55.055Z"}]}