NGL AI Moderation: Harm-Filtering Claim Evidence Questions
The FTC and Los Angeles District Attorney's Office alleged NGL falsely claimed its AI content moderation filtered cyberbullying and other harmful messages. This case shows the evidence burden behind AI safety claims for apps used by children or teens.
Source update, company response, or correction? Send a private note for review →
What was claimed
NGL marketed its anonymous messaging app as a safer place for teens and claimed it used world-class AI content moderation, including deep learning and pattern matching algorithms, to combat cyberbullying and other harms.
- Risk pattern
- AI harm-filtering claim without disclosed detection scope, error rates, or child-safety limits
Why this mattered
AI moderation claims are safety claims. Buyers, parents, platform partners, or users need to know what categories are filtered, what the missed-message rate is, what false positives occur, how appeals or escalation work, and whether protections hold for younger users, slang, coded language, images, links, or repeated harassment patterns.
What the source said
The FTC press release said NGL and its co-founders falsely claimed the app's AI content moderation program filtered cyberbullying and other harmful messages. It also said the company marketed to kids and teens, sent fake messages that appeared to come from real people, and would be prohibited from misrepresenting AI capabilities.
Evidence gap / buyer questions
Harm category definitions, training data and evaluation set, false negative and false positive rates, teen-user test coverage, language and slang coverage, escalation process, human review boundary, recurrence handling for repeat abuse, and update process for new harmful-message patterns.
- Which harmful-message categories does the AI moderation claim cover, and which categories are out of scope?
- What false negative rate appears for bullying, threats, harassment, sexual content, or coded language?
- When does a human reviewer inspect messages that the model misses or flags incorrectly?
- How is moderation performance tested for teen users, slang, multilingual messages, repeated abuse patterns, and anonymous-message prompts?
How this applies to your vendor evaluation
If a vendor you are evaluating makes a claim with this pattern, copy the exact sentence and review that wording against the evidence standard this case documents.
Wording boundary direction
Uses automated moderation to flag selected message categories for review; coverage, false negative rates, and escalation limits are documented at [link].
A lower-risk wording boundary narrows the scope, discloses the test conditions, and does not overstate what is covered.
Update and response status
Disclaimer / correction note
This case description draws from the FTC source cited above. It is not legal advice, a child-safety certification, or a moderation-system audit.
This tool generates evidence-burden notes, evidence requests, and buyer questions based on publicly accessible source content. It does not determine whether a product is true, false, compliant, or suitable for any purpose. It is not legal, investment, procurement, or professional compliance advice. See the full disclaimer.
Source update, company response, or correction? Submit a correction or source note →
Check a similar public AI claim
Review the exact public URL or wording for evidence gaps and buyer questions.