Appearance
Risk type — the full vocabulary
Recorded vocabulary snapshot. This page is generated from a saved database extract, not a live connection. Rebuilding the site does not refresh its measurements. Draft wording comes from review notes; check recorded decisions for approved meaning. Stored values alone do not prove current prompt or filter behavior.
There are 30 registered values, and 92 more words in use that are not registered.
Where it is stored: content_risk_signals.risk_type — 4,644 rows carry a value.
How the registry and the data line up: 30 of the 30 registered values appear in the data, and 92 words appear that are not registered.
What this describes
— no definition written yet.
The values
| Value | Stored as | In use | Recorded wording |
|---|---|---|---|
| Accuracy | accuracy | 643 | — |
| Product Confusion | product_confusion | 575 | — |
| Capability Mismatch | capability_mismatch | 566 | — |
| Performance-Claim Skepticism | perf_claim_skepticism | 534 | — |
| Overpromising | overpromising | 432 | — |
| Safety | safety | 321 | — |
| Privacy | privacy | 281 | — |
| Price / Value Skepticism | price_value_skepticism | 169 | — |
| Security | security | 145 | — |
| AI Replacing Human Feeling | ai_replacing_human_feeling | 143 | — |
| Regulatory | regulatory | 142 | — |
| Creator Replacement | creator_replacement | 128 | — |
| Deepfakes | deepfakes | 73 | — |
| Bias / Fairness | bias_fairness | 64 | — |
| Copyright | copyright | 49 | — |
| Cultural Misread | cultural_misread | 48 | — |
| Children / Family | children_family | 43 | — |
| Enterprise Data Leakage | enterprise_data_leakage | 40 | — |
| Hallucination | hallucination | 31 | — |
| Misuse | misuse | 19 | — |
| Too Technical | too_technical | 19 | — |
| Publisher Traffic Loss | publisher_traffic_loss | 19 | — |
| Control | control | 14 | — |
| Quality | quality | 12 | — |
| Misinformation | misinformation | 8 | — |
| Trust | trust | 7 | — |
| Prompt Injection | prompt_injection | 4 | AI-security risk: adversarial instructions embedded in content/inputs. |
| Trust & Safety | trust_safety | 4 | Platform trust-and-safety domain risk (moderation, abuse, T&S posture) - distinct from brand 'trust' and product 'safety'. |
| Greenwashing | greenwashing | 3 | — |
| Luxury Dilution | luxury_dilution | 1 | — |
Words in use that nobody registered
Each of these was written into the data without a registry row behind it. A word that is not registered cannot be offered as a choice, defined, or counted on.
| Word | Rows carrying it |
|---|---|
differentiation | 3 |
disinformation | 3 |
ecosystem | 3 |
conflict_of_interest | 2 |
dual_use_risks | 2 |
economic_inequality | 2 |
liability | 2 |
national_security | 2 |
societal_disruption | 2 |
societal_impact | 2 |
transparency | 2 |
user_experience | 2 |
abuse | 1 |
access_limitation_frustration | 1 |
account_recovery | 1 |
account_security | 1 |
adaptation | 1 |
adversarial_attacks | 1 |
ai_reputational_harm | 1 |
alignment_tax | 1 |
anthropomorphization | 1 |
autonomous_weapons | 1 |
autonomy_agency | 1 |
biorisk | 1 |
bugs_bad_behavior | 1 |
cognitive_atrophy | 1 |
competitive_positioning | 1 |
competitive_positioning_risk | 1 |
complexity | 1 |
compliance | 1 |
compute_requirements | 1 |
concentration_of_power | 1 |
concentration_of_wealth_power | 1 |
connectivity | 1 |
contamination | 1 |
context_management | 1 |
cybersecurity_risk | 1 |
data_privacy | 1 |
data_quality_issues | 1 |
deception | 1 |
disruption | 1 |
domestic_surveillance | 1 |
doomscrolling | 1 |
economic_displacement | 1 |
emotional_distress | 1 |
emotional_support | 1 |
entropy | 1 |
ethical_challenges | 1 |
ethical_concerns | 1 |
execution_risk | 1 |
existential_risk | 1 |
frontier_risks | 1 |
geopolitical_competition | 1 |
harmful_content | 1 |
high_stakes_decisions | 1 |
impersonation | 1 |
inconsistent_behavior | 1 |
information_hazards | 1 |
intense_competition | 1 |
investment_return_risk | 1 |
jailbreaks | 1 |
jobs_economy | 1 |
labor_shortages | 1 |
lock_in | 1 |
misuse_potential | 1 |
model_autonomy | 1 |
model_mistakes | 1 |
organizational_change_resistance | 1 |
oversimplification | 1 |
persuasion | 1 |
phishing | 1 |
productivity | 1 |
reliability | 1 |
reputation_risk | 1 |
responsible_ai | 1 |
reward_hacking | 1 |
scalability | 1 |
scams | 1 |
shadow_ai | 1 |
sustainability | 1 |
sycophancy | 1 |
technical_complexity | 1 |
token_usage | 1 |
too_vague | 1 |
toxicity | 1 |
trust_proof | 1 |
unauthorized_voice_generation | 1 |
uncertainty | 1 |
unemployment_shocks | 1 |
unreliability | 1 |
vendor_risk | 1 |
wellbeing_and_dependency | 1 |
How complete this page is
Of 30 registered values, 2 carry a written definition (7%) and 0 carry tagging rules (0%). These counts are calculated from the saved database extract. They describe that snapshot, not a fresh database check.
Where to look next
- What this is FOR, in plain language — the Content chapter.
- Every other vocabulary — the vocabulary index.
- What is still undecided about it — the open questions.
- The exact tables and columns — the data reference.