Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
---
|
| 2 |
-
title: Dark Pattern
|
| 3 |
-
emoji:
|
| 4 |
colorFrom: purple
|
| 5 |
colorTo: red
|
| 6 |
sdk: static
|
|
@@ -10,22 +10,32 @@ tags:
|
|
| 10 |
- nlp
|
| 11 |
- consumer-protection
|
| 12 |
- distilbert
|
| 13 |
-
- ethics-in-ai
|
| 14 |
---
|
| 15 |
|
| 16 |
-
#
|
| 17 |
|
| 18 |
-
|
| 19 |
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
2. π‘ **Misdirection:** Pre-selected add-ons and deceptive visual hierarchy.
|
| 23 |
-
3. π£ **Social Proof:** Synthetic activity badges ("50 people looking right now").
|
| 24 |
-
4. π©· **Forced Continuity:** Friction-heavy cancellation and disguised recurring renewals.
|
| 25 |
-
5. π **Confirm-shaming:** Manipulative opt-out copy designed to guilt-trip users.
|
| 26 |
-
6. π©΅ **Hidden Costs:** Undisclosed mandatory charges revealed only at final payment.
|
| 27 |
-
7. π΅ **Disguised Ads:** Sponsored placement camouflaged as organic editorial content.
|
| 28 |
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Dark Pattern Detector
|
| 3 |
+
emoji: π‘οΈ
|
| 4 |
colorFrom: purple
|
| 5 |
colorTo: red
|
| 6 |
sdk: static
|
|
|
|
| 10 |
- nlp
|
| 11 |
- consumer-protection
|
| 12 |
- distilbert
|
|
|
|
| 13 |
---
|
| 14 |
|
| 15 |
+
# Dark Pattern Detector π‘οΈ
|
| 16 |
|
| 17 |
+
An interactive NLP classification engine designed to detect deceptive UI copy, artificial scarcity, and manipulative e-commerce patterns.
|
| 18 |
|
| 19 |
+
- **GitHub Repository**: [JayanGupta/Dark-Pattern-Detection](https://github.com/JayanGupta/Dark-Pattern-Detection)
|
| 20 |
+
- **Author**: [Jayan Gupta](https://huggingface.co/Jayan12321)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
+
---
|
| 23 |
+
|
| 24 |
+
## π― What are Dark Patterns?
|
| 25 |
+
Dark patterns are user interface designs crafted to trick users into doing things they might not otherwise do (e.g. buying unwanted add-ons, signing up for disguised recurring subscriptions, or agreeing to invasive tracking).
|
| 26 |
+
|
| 27 |
+
---
|
| 28 |
+
|
| 29 |
+
## π Detected Taxonomies
|
| 30 |
+
This tool identifies manipulative patterns across industry-standard taxonomies:
|
| 31 |
+
1. **Urgency & Scarcity**: Artificial countdown timers and fake low-stock alerts.
|
| 32 |
+
2. **Misdirection**: Using visual hierarchy or pre-checked checkboxes to steer users toward a specific choice.
|
| 33 |
+
3. **Social Proof Exploitation**: Synthetic activity notifications ("34 people looking at this right now").
|
| 34 |
+
4. **Forced Continuity**: Easy one-click signups combined with labyrinthine cancellation flows.
|
| 35 |
+
5. **Confirm-shaming**: Emotionally manipulative copy on opt-out buttons ("No thanks, I hate saving money").
|
| 36 |
+
6. **Hidden Costs**: Unannounced fees revealed only on the final checkout screen.
|
| 37 |
+
|
| 38 |
+
---
|
| 39 |
+
|
| 40 |
+
## π οΈ Development & Transparency Note
|
| 41 |
+
Designed and developed by **Jayan Gupta**. The taxonomy and heuristics are based on FTC consumer protection guidelines and academic dark-pattern datasets. AI assistants were used for prototyping and documentation structuring.
|