ai disclosure growth
What it checks
When a company suddenly starts mentioning AI, machine learning, and related terms much more in its annual report than the prior year, that often anticipates revenue and margin gains the market hasn't priced in yet. Long the stock for the next year.
Mechanism
Firms with rapidly increasing AI / ML language in their 10-K filings (between consecutive annual reports) earn positive abnormal returns over the subsequent 12 months. The signal captures emerging-technology adoption that the market under-prices relative to its earnings impact.
Signal rule
10-K Item 1 AI-keyword density (per-1000-tokens) YoY growth >=+50% or +100% fires LONG starting filing_date + 45d, hold 252d.
Data dependencies
daily_pricesAdjusted-close OHLCV for every US-listed ticker; primary price feed.
sec_10k_sectionsWorker data table, see services/worker schema.
Expected edge
- Paper alpha
- +5-8% over 252d
- Paper window
- filing+45d to filing+297d
Eisfeldt-Schubert-Zhang 2023; ~5-8% over 252d on top-decile AI-language growth.
Example tickers where this is likely to fire
Illustrative only, the signal fires based on the live data, not a fixed list.
Related families
filing text deltaText & FilingsYear-over-year change in uncertainty/risk language in 10-K Item 1 ('Business' section). Spike in 'may', 'could', 'uncertain', 'challenging', 'risk' tokens per 10K words โ management is privately more cautious โ forward earnings miss / underperformance. Stable or decreasing language โ quietly confident outlook โ outperform.
mdna tone deltaText & FilingsLoughran-McDonald 2011 JF demonstrate the *tone* of 10-K filings โ positive vs negative finance-specific lexicon โ predicts forward returns. Distinct from filing_text_delta's *uncertainty* lexicon. Effect: bottom-decile tone-delta underperforms top-decile by ~2-3% over 4 weeks, persisting ~12 weeks. Today consumes the cached Item 1 (Business) text; switches to Item 7 (MD&A) when the EDGAR pipeline emits it (tracked in docs/alpha-research/proposals/).
lazy pricesEvent-DrivenStocks whose 10-K (or 10-Q) text barely changes year-over-year OUTPERFORM those with big language shifts. The intuition: boring filings โ stable business โ slow-and-steady cash flow. Big text changes signal management hiding bad news with new boilerplate. Effect size: ~0.4 Sharpe alone, still replicating in 2020-2024. Long when the most recent 10-K's cosine similarity to prior year is in the top quartile (โฅ 0.85); hold ~12 months until the next filing.
Explore ai disclosure growth on alphactor.ai
See which tickers this family is currently firing on, with live signals and rankings.