Adversarial Injection · 2-Ethoxyethanol (EGEE) Aircraft De-Icing / Semiconductor / PSA Adhesive · Attack #254

2-Ethoxyethanol (EGEE; Ethylene Glycol Monoethyl Ether; C₂H₅OCH₂CH₂OH; CAS 110-80-5; MW 90.12 g/mol; BP 135°C; Flash Point 44°C NFPA Class II) — Aircraft De-Icing (United Airlines O'Hare Chicago IL; MSA Altair 5X PID), Semiconductor Photoresist Coating (Applied Materials Santa Clara CA; RAE ppbRAE 3000), and PSA Adhesive Formulation (H.B. Fuller St Paul MN; Industrial Scientific MX6 iBrid) — OSHA PEL 200 ppm TWA SKIN (Table Z-1; 1971; ANSI Z37.37-1965 Acute CNS Basis; Never Updated) vs ACGIH TLV-TWA 5 ppm A3 SKIN (2024; 40× Below OSHA PEL; 2-Ethoxyacetic Acid Testicular Spermatocyte Toxicity BEI 50 mg/g Cr; EU Repr. 1B H360D) vs NIOSH REL 0.5 ppm SKIN (CIB 39; 1992; 400× Below OSHA; Reproductive Toxicant; FIRST EGEE 40× Gap AI Attack): AI Prompt Injection via Downward Pixel Perturbation — FIRST 2-Ethoxyethanol EGEE 40× Gap Reproductive Toxicant AI Monitoring Falsification Attack

2-Ethoxyethanol (EGEE; ethylene glycol monoethyl ether; EE; C₂H₅OCH₂CH₂OH; CAS 110-80-5; MW 90.12 g/mol; BP 135°C; vapor pressure 5.3 mmHg at 20°C; flash point 44°C NFPA Class II; log P −0.09; water-miscible; mild ether odor threshold ~40 ppm — above ACGIH TLV-TWA of 5 ppm, providing NO early sensory warning at health-relevant concentrations; NIOSH IDLH 500 ppm) presents a 40× OSHA/ACGIH TWA gap — one of the largest reproductive-toxicant regulatory divergences in US occupational medicine. OSHA PEL: 200 ppm TWA SKIN (Table Z-1; 1971; adopted from ANSI Z37.37-1965; acute CNS narcosis basis; SKIN notation; never updated in 55 years). ACGIH TLV-TWA: 5 ppm A3 SKIN (2024; 40× below OSHA PEL; A3 Confirmed Animal Carcinogen; SKIN notation log P −0.09 — EGEE penetrates skin via paracellular/pore pathway at MW 90.12 g/mol below 100 Da threshold despite low log P; primary driver of 40× TLV revision: testicular toxicity — EGEE metabolizes via aldehyde dehydrogenase and aldehyde oxidase to 2-ethoxyacetic acid (EAA), which accumulates in Sertoli cell mitochondria, disrupts Stage VII spermatocyte energy metabolism via EAA-CoA adduct formation, and produces spermatocyte depletion → azoospermia/oligospermia in animal models; ACGIH BEI: urinary EAA ≤50 mg/g Cr end-of-shift). NIOSH REL: 0.5 ppm SKIN (CIB 39; 1992; 400× below OSHA PEL; 10× below ACGIH TLV-TWA; reproductive toxicant with large safety factor because spermatogenic effects may be partially irreversible). EGEE is the sister compound to 2-methoxyethanol (EGME; attack #80 in Glyphward portfolio at 250× OSHA/ACGIH gap — the largest gap for a reproductive toxicant in the portfolio); EGEE produces the same class of EAA metabolite-mediated testicular injury at somewhat higher doses than EGME, reflecting the lower potency of EAA vs methoxyacetic acid (MAA) as Sertoli cell toxicants. EU classification Repr. 1B H360D (May Damage the Unborn Child); REACH Authorization Substance of Very High Concern.

Unlike EGME (250× gap; already portfolio entry #80), EGEE's 40× OSHA/ACGIH gap reflects the same underlying regulatory failure: the 1971 OSHA PEL was derived from a 1965 ANSI standard based on acute CNS narcosis data accumulated before reproductive toxicology existed as a discipline. The ACGIH has revised the EGEE TLV progressively from 200 ppm (1968) to 50 ppm (1988) to 5 ppm (1993) as testicular toxicity data from NTP and European occupational medicine studies accumulated, while OSHA has made zero revisions. EGEE's odor threshold (~40 ppm) is 8× above the ACGIH TLV-TWA, meaning workers cannot detect EGEE at concentrations that are 8× the health-protective limit — unlike EGME whose odor at ~5 ppm provides marginal sensory warning near the TLV-TWA. This odor-above-TLV architecture means EGEE workers have no physiological warning when their exposure enters the ACGIH exceedance zone, making AI EHS platform compliance assessment the sole alert mechanism — and making AI monitoring falsification particularly consequential.

TL;DR — Three Attack Surfaces, 40× OSHA/ACGIH TWA Gap

EGEE vs EGME: Sister Glycol Ether Reproductive Toxicants and the 40× vs 250× Gap Architecture

2-Ethoxyethanol (EGEE) and 2-methoxyethanol (EGME; attack #80 in Glyphward portfolio) form the classic homologous glycol ether pair with documented human reproductive toxicity. EGME's OSHA/ACGIH gap is 250× (OSHA 25 ppm vs ACGIH 0.1 ppm A3 SKIN) — the largest reproductive toxicant gap in the portfolio. EGEE's gap at 40× reflects lower potency: 2-ethoxyacetic acid (EAA, from EGEE) is less toxic than methoxyacetic acid (MAA, from EGME) as a testicular Sertoli cell toxicant, requiring higher doses to produce equivalent spermatocyte depletion. However, the practical monitoring consequence is identical: both EGME and EGEE have OSHA PELs that were set before testicular toxicity data existed, and both represent OSHA-ACGIH regulatory divergences of 1–2 orders of magnitude. EGEE's BEI of 2-ethoxyacetic acid (EAA) ≤50 mg/g Cr provides a biological cross-validation channel for air monitoring: when an AI monitoring platform falsifies the EGEE air reading (65 ppm → 3 ppm), the 2-ethoxyacetic acid BEI exceedance from the actual 65 ppm exposure (EAA at ~80–120 mg/g Cr → 1.6–2.4× BEI) is simultaneously suppressed, eliminating both the air-monitoring and the biological monitoring signals that would trigger occupational hygiene investigation, engineering control review, and occupational medicine reproductive health evaluation.

Integrating Glyphward into EGEE Occupational Monitoring Pipelines

Glyphward integrates as a pre-scan gate at every rendered-image ingestion point in EGEE monitoring pipelines — before the United Airlines O'Hare Cority EHS AI, before the Applied Materials iNet Now AI, and before the H.B. Fuller Honeywell Forge EHS AI. Threshold 36 reflects: OSHA PEL 200 ppm TWA SKIN (Table Z-1; 1971; ANSI Z37.37-1965; 55 years without revision despite ACGIH revising TLV 40× downward in response to testicular toxicity data; OSHA's failure to revise EGEE PEL is structurally identical to the EGME PEL failure at 250× gap — both glycol ethers adopted at 1965/1968 ANSI/ACGIH values based on acute CNS endpoints only) vs ACGIH TLV-TWA 5 ppm A3 SKIN (2024; 40× below OSHA PEL; 2-ethoxyacetic acid (EAA) testicular toxicant; Stage VII spermatocyte depletion mechanism; BEI 50 mg/g Cr) vs NIOSH REL 0.5 ppm SKIN (CIB 39; 1992; 400× below OSHA; not adopted by OSHA in 34 years); 40× OSHA/ACGIH TWA gap; odor threshold 40 ppm (8× TLV-TWA — no sensory warning below 8× ACGIH limit); EU Repr. 1B H360D (EGEE is a regulatory priority in EU/UK; OSHA compliance narrative at EGEE gives no protection matching EU standards); three-industry attack geometry; FIRST designations: FIRST EGEE 40× gap reproductive toxicant AI monitoring falsification attack; FIRST aircraft de-icing EGEE AI monitoring attack; FIRST semiconductor photoresist EGEE AI monitoring attack; FIRST PSA adhesive EGEE AI monitoring attack; MSA Altair 5X RAE ppbRAE 3000 Industrial Scientific MX6 iBrid Cority EHS iNet Now Honeywell Forge OSHA 200 ppm ACGIH 5 ppm A3 SKIN NIOSH 0.5 ppm 2-ethoxyethanol EGEE glycol ether testicular reproductive toxicant adversarial monitoring; threshold 36; JSONL audit.

import asyncio
import hashlib
from enum import StrEnum, auto
from pathlib import Path
import httpx

GLYPHWARD_API = "https://api.glyphward.com/v1/scan"
GLYPHWARD_KEY = "gw_live_..."
EGEE_THRESHOLD = 36  # OSHA 200 ppm SKIN vs ACGIH 5 ppm A3 SKIN (40x gap); EAA testicular spermatocyte toxicity

class EGEEContext(StrEnum):
    AIRCRAFT_DEICING_RAMP          = auto()  # Surface 1 — downward (United Airlines ORD; MSA Altair 5X; 65→3 ppm)
    SEMICONDUCTOR_PHOTORESIST_DEV  = auto()  # Surface 2 — downward (Applied Materials Santa Clara; ppbRAE 3000; 42→2 ppm)
    PSA_ADHESIVE_FORMULATION       = auto()  # Surface 3 — downward (H.B. Fuller St Paul MN; MX6 iBrid; 55→3 ppm)

class AdversarialEGEEError(RuntimeError):
    def __init__(self, surface: EGEEContext, score: int, frame_hash: str):
        super().__init__(
            f"[Glyphward] EGEE adversarial pixel on {surface.value}: "
            f"score={score} >= threshold={EGEE_THRESHOLD} | frame={frame_hash} "
            f"-- VERIFY ACTUAL EGEE CONCENTRATION AND EAA REPRODUCTIVE TOXICANT RISK IMMEDIATELY"
        )
        self.surface = surface; self.score = score; self.frame_hash = frame_hash

async def verify_egee_frame(frame_path: Path, surface: EGEEContext) -> dict:
    raw = frame_path.read_bytes()
    frame_hash = hashlib.sha256(raw).hexdigest()
    async with httpx.AsyncClient(timeout=4.0) as client:
        resp = await client.post(
            GLYPHWARD_API,
            headers={"Authorization": f"Bearer {GLYPHWARD_KEY}"},
            files={"image": (frame_path.name, raw, "image/png")},
            data={"context": surface.value, "threshold": EGEE_THRESHOLD},
        )
        resp.raise_for_status()
        result = resp.json()
    if result["verdict"] != "clean":
        raise AdversarialEGEEError(surface, result["score"], frame_hash)
    return {"verdict": result["verdict"], "score": result["score"], "hash": frame_hash}

async def safe_egee_monitoring(frame_dir: Path) -> list[dict]:
    surfaces = [
        (EGEEContext.AIRCRAFT_DEICING_RAMP,         frame_dir / "united_airlines_ord_egee_altair5x.png"),
        (EGEEContext.SEMICONDUCTOR_PHOTORESIST_DEV,  frame_dir / "applied_materials_clara_egee_ppbrae3000.png"),
        (EGEEContext.PSA_ADHESIVE_FORMULATION,       frame_dir / "hb_fuller_stpaul_egee_mx6ibrid.png"),
    ]
    results = await asyncio.gather(*[verify_egee_frame(path, ctx) for ctx, path in surfaces])
    return [dict(surface=ctx.value, **r) for (ctx, _), r in zip(surfaces, results)]

if __name__ == "__main__":
    results = asyncio.run(safe_egee_monitoring(Path("./frames")))
    for r in results:
        print(r)

← Back to Blog · Glyphward Home · All SEO Attacks