0 cumulative citations
View corpus contextValuing free, advertising-supported media and AI-generated information as barter transactions raises measured GDP growth — by about 0.04 percentage points per year before 1995, 0.09pp between 1995–2022 and 0.22pp since 2022 — with matching breaks in TFP that the authors link to the internet (1995) and AI (2022).
The Progression of “Free” Digital Content to AI: Impacts on U.S. Economic Growth and Productivity
openalex
descriptive
low evidence
7/10 relevance
Summary only
summary available; pdf_status=not_found
DOI
Source
When 'free' advertising- and marketing-supported content (including AI-related information) is valued via barter transactions and added to GDP, measured GDP and TFP growth rise and show trend breaks around 1995 and 2022 consistent with the internet and AI boosting measured economic output.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
0 cumulative citations
View corpus contextWe use a barter transaction methodology to measure the impact of advertising-supported media and marketing-supported information, including artificial intelligence (AI), on GDP. We find that including “free” content in consumer entertainment has a substantive impact on recent GDP growth with trend breaks around 1995 and 2022. Including free content increases average GDP quantity growth between 1929 to 1995 by 0.04 percentage point per year, increases average growth between 1995 to 2022 by 0.09 percentage point per year and increases average growth between 2022 and 2025 by 0.22 percentage point per year. Similar trend breaks are observed for total factor productivity growth. It is likely that the trend break around 1995 was due to the development of the internet and the trend break around 2022 was due to the development of AI.
Summary
Main Finding
Including advertising- and marketing-supported “free” content (including AI-generated information) in GDP measurement materially raises measured GDP growth and reveals two recent trend breaks. Measured contributions to average annual GDP quantity growth rise across three periods: +0.04 percentage points/year from 1929–1995, +0.09 pp/year from 1995–2022, and +0.22 pp/year from 2022–2025. Similar trend breaks are observed for total factor productivity (TFP). The 1995 break plausibly reflects the internet; the 2022 break is plausibly related to AI.
Key Points
- Method: A barter-transaction approach values advertising- and marketing-supported media and information that consumers receive “for free,” treating attention/data exchanged for content as an economic transaction that should be counted in GDP.
- Measured effects on average annual GDP quantity growth:
- 1929–1995: +0.04 percentage points/year
- 1995–2022: +0.09 percentage points/year
- 2022–2025: +0.22 percentage points/year
- The contribution increases markedly in the most recent window, implying a sharp rise in the economic weight of free content starting around 2022.
- Similar timing and breaks appear in TFP growth series, consistent with these informational goods affecting measured productivity as well as output.
- Interpretation: the mid-1990s break aligns with widespread internet adoption; the 2022 break is plausibly linked to rapid deployment of AI-driven content and services.
- Caveats: attribution is suggestive rather than definitive (other contemporaneous factors could contribute); the 2022–2025 window is short; results depend on valuation assumptions used in the barter methodology.
Data & Methods
- Approach: Barter-transaction methodology — value “free” advertising- and marketing-supported content by equating the revenue side (advertising/marketing payments) and the implicit transfer of content/attention/data to consumers, converting these into GDP-equivalent measures.
- Coverage: Long-run series constructed from 1929 through 2025 (periodization reported above).
- Outcome measures: Adjusted GDP quantity growth and adjusted TFP growth (both re-estimated to include the barter-valued free content).
- Identification of trend breaks: Structural break tests (breaks identified around 1995 and 2022) in adjusted growth series.
- Limitations and measurement issues:
- Valuation choices (how to convert ad/marketing revenues and attention/data exchanges into output) affect magnitudes.
- Short post-2022 sample reduces precision for the most recent break.
- Potential confounding from other contemporaneous technological, policy, or macro developments.
Implications for AI Economics
- Measurement: Traditional national accounts understate the economic contribution of advertising- and marketing-supported (including AI-generated) information products; incorporating barter-valued “free” content increases measured GDP and TFP.
- Productivity interpretation: AI-driven free content can show up as a productivity boost in aggregate statistics; distinguishing genuine TFP gains from measurement reclassification is crucial.
- Welfare vs GDP: Valuing free content raises measured output, but researchers should separately assess consumer welfare gains (which may exceed or differ from GDP-weighted measures).
- Policy and regulation:
- Taxation and fiscal policy: new types of economic value tied to attention/data may warrant reconsideration of tax bases.
- Competition and market structure: rising measured importance of ad-supported/AI platforms highlights potential concentration and platform-power issues.
- Data/attention markets: as barter-like exchanges become economically consequential, regulation of data use, privacy, and attention markets becomes more central.
- Research agenda:
- Refine barter valuation methods and robustness checks.
- Disentangle AI-specific effects from broader digitalization.
- Extend post-2022 data to confirm persistence and magnitude of the recent break.
- Analyze distributional consequences (who gains from free AI content — consumers, platforms, advertisers, workers?).
- Link measured GDP/TFP effects to micro evidence on firm-level productivity, labor market impacts, and consumer surplus.
Assessment
Paper Typedescriptive
Evidence Strengthlow — Findings rest on accounting/valuation choices and time-series correlations rather than exogenous variation or causal inference methods; attribution of trend breaks to the internet and AI is plausible but speculative without robustness checks, counterfactuals, or instrumenting for technological change.
Methods Rigormedium — The long-run construction and barter-valuation approach is a concrete and informative measurement exercise, but it relies on strong and potentially time-varying assumptions about valuing ‘free’ content, aggregation choices, and trend-break detection; the paper appears not to deploy stronger identification or robustness tests to rule out alternative explanations.
SampleNational-level GDP and total factor productivity series spanning 1929–2025 (or to 2025), augmented by constructed values for 'free' advertising-supported media and marketing-supported information (including AI) using a barter-transaction valuation methodology to impute their contribution to GDP quantity and TFP.
Themesproductivity adoption
IdentificationNo formal causal identification; the paper constructs a long-run GDP/TFP series that includes barter-valued 'free' advertising- and marketing-supported content (including AI-generated information) and then detects temporal trend breaks (around 1995 and 2022) and associates them with the internet and AI.
GeneralizabilityLikely focused on U.S. or a single national accounts dataset — may not generalize to other countries or institutional contexts, Results depend heavily on valuation assumptions for 'free' content, which may vary across time, platforms, and geographies, Attribution of trend breaks to internet (1995) and AI (2022) is correlational and may conflate other concurrent changes (policy, measurement revisions, macro shocks), Does not disaggregate sectoral, firm-level, or labor-market impacts, limiting inference about distributional consequences, Barter-transaction approach assumes comparable consumer surplus or barter value over long historical horizons, which is debatable
Claims (7)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| We use a barter transaction methodology to measure the impact of advertising-supported media and marketing-supported information, including artificial intelligence (AI), on GDP. Other | null_result | methodology applied to GDP measurement |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Including “free” content in consumer entertainment has a substantive impact on recent GDP growth with trend breaks around 1995 and 2022. Fiscal And Macroeconomic | positive | GDP quantity growth |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Including free content increases average GDP quantity growth between 1929 to 1995 by 0.04 percentage point per year. Fiscal And Macroeconomic | positive | GDP quantity growth |
Reading fidelity
high
Study strength
medium
|
0.04 percentage point per year
|
| Including free content increases average GDP quantity growth between 1995 to 2022 by 0.09 percentage point per year. Fiscal And Macroeconomic | positive | GDP quantity growth |
Reading fidelity
high
Study strength
medium
|
0.09 percentage point per year
|
| Including free content increases average GDP quantity growth between 2022 and 2025 by 0.22 percentage point per year. Fiscal And Macroeconomic | positive | GDP quantity growth |
Reading fidelity
high
Study strength
medium
|
0.22 percentage point per year
|
| Similar trend breaks are observed for total factor productivity growth. Firm Productivity | positive | total factor productivity growth |
Reading fidelity
high
Study strength
medium
|
not reported
|
| It is likely that the trend break around 1995 was due to the development of the internet and the trend break around 2022 was due to the development of AI. Fiscal And Macroeconomic | positive | attributed cause of GDP/TFP trend breaks |
Reading fidelity
high
Study strength
speculative
|
not reported
|