#I received an email stating that content I generated may violate the use policy, and I'm unsure wha

1 messages · Page 1 of 1 (latest)

feral krakenBOT
#
AI-generated Answer for goofusmax_54890

Short answer: both. ElevenLabs still blocks some requests before generation, and it also enforces policy after the fact.

What’s proactive today

  • Automated moderation scans inputs and can “block [violations] outright or flag them for review.” This is platform‑wide, not limited to a single product. (elevenlabs.io)
  • Voice Library “Live Moderation” (if the voice owner enabled it) checks the text before audio is generated and may block requests; it can add some latency. (help.elevenlabs.io)
  • “No‑go voices” proactively prevent creation/use of certain high‑risk voice clones (e.g., political figures). (help.elevenlabs.io)

What’s retroactive

  • Policy enforcement also happens after generation using “automated systems, user reports, and human review,” and may result in content removal or account actions. (elevenlabs.io)
  • Terms note ElevenLabs may refuse to post or remove content “at any time and for any reason with or without notice,” so you might not always see an in‑product warning even if an email is sent. (elevenlabs.io)

How to investigate your case

  • Review your account History to see recent generations that might relate to the email. (help.elevenlabs.io)
  • If you still can’t identify the item, contact Support via the help‑center form with timestamps and any relevant details from the email. (help.elevenlabs.io)

References: Safety overview; Prohibited Use Policy; Terms; Live Moderation; No‑go voices; History; Contact Support. (elevenlabs.io, help.elevenlabs.io)

Sources:
ElevenLabs — Safety | ElevenLabs
What is Live Moderation? – ElevenLabs
What are No-Go Voices? – ElevenLabs
ElevenLabs Prohibited Use Policy | ElevenLabs
ElevenLabs — Terms of Use EU | ElevenLabs
Can I access the files I generated in the past? – ElevenLabs
How can I contact support? – ElevenLabs
ElevenLabs — Safety | ElevenLabs
What is Live Moderation? – ElevenLabs

wicked plover
#

Okay. Though, I thought that live moderation was only used on professional clones, which I have not used any. It seems like the content was subject to live moderation even though I can't seem to see anything that says it should have been, and I don't know what content was the source of the problem either.

feral krakenBOT
#
AI-generated Answer for goofusmax_54890

Short answer:

  • Live Moderation tied to Professional Voice Clones exists, but it isn’t the only moderation we use. If you didn’t use a professional clone, your content could still be checked by our platform-wide automated/human moderation or (if you’re using ConvAI/agents) message‑level live moderation. (help.elevenlabs.io, elevenlabs.io)

Details:

  • Live Moderation (Voice Library): This is an optional setting the voice owner can enable when sharing a Professional Voice Clone via the Voice Library; it checks text for prohibited categories and can add latency. It only applies when using that shared voice with Live Moderation enabled. (help.elevenlabs.io)
  • Platform-wide moderation: Separately from the above, our safety page explains we “actively monitor content,” using automated systems to scan content and either block it outright or flag it for review (which can result in retroactive action). This applies across our services, not only to professional clones. (elevenlabs.io)
  • ConvAI message-level live moderation: For voice agents, message-level live moderation can be enabled at the workspace level (enabled by default in some cases). It drops a response if prohibited content is detected before it’s spoken; currently, the blog specifies SCIM-related content is blocked, with minimal added latency. (elevenlabs.io)

What this means for your case:

  • Yes, prompts/outputs can be restricted before generation (blocked), or addressed after the fact if flagged by automated systems or human review—both mechanisms exist. (elevenlabs.io)
  • If you’re unsure which item triggered the email, you can review your History (all generated data is accessible in the History tab), and you can contact Support for specifics or to dispute an action. (help.elevenlabs.io)

References to policies/categories:

  • See the Prohibited Use Policy for the categories we enforce against. (elevenlabs.io)

Sources:
What is Live Moderation? – ElevenLabs
ElevenLabs — Safety | ElevenLabs
What is Live Moderation? – ElevenLabs
ElevenLabs — Safety | ElevenLabs
An overview of Safety framework for AI voice agents | ElevenLabs
ElevenLabs — Safety | ElevenLabs
How do I make a Data Subject Request? – ElevenLabs
ElevenLabs Prohibited Use Policy | ElevenLabs

wicked plover
#

What is the difference between live moderation and the generalized automated moderation?

feral krakenBOT
#
AI-generated Answer for goofusmax_54890

Short answer:

  • “Live Moderation” (capitalized) is a specific, opt‑in pre‑generation check tied to Professional Voice Clones shared via the Voice Library. When the voice owner enables it, incoming text requests for that voice are screened against prohibited categories before audio is generated; this can add latency. (help.elevenlabs.io)
  • “Automated moderation” (platform‑wide) refers to ElevenLabs’ general safety systems that continuously scan content across the service. These systems may block content outright or flag it for human review, and enforcement actions (e.g., removing material or suspending access) can occur proactively or after generation. The docs do not use the phrase “generalized automated moderation,” but describe “automated moderation” plus human review. (elevenlabs.io)

Related note:

  • Conversational AI also supports “message‑level live moderation” that can drop a call if a prohibited response is detected; this is configured at the workspace/agent level and is separate from the Voice Library setting. (elevenlabs.io)

Where they apply:

  • Live Moderation: only on Voice Library Professional Voice Clones whose owners enabled it. (help.elevenlabs.io)
  • Automated moderation: applies broadly across ElevenLabs products and accounts, independent of whether a specific voice has Live Moderation enabled. (elevenlabs.io)

Sources:
What is Live Moderation? – ElevenLabs
ElevenLabs — Safety | ElevenLabs
An overview of Safety framework for AI voice agents | ElevenLabs
What is Live Moderation? – ElevenLabs
ElevenLabs — Safety | ElevenLabs

wicked plover
#

Looking back through my generation history there was some dialogue that was slightly sexual, is content like that prohibited under all use cases?

feral krakenBOT
#
AI-generated Answer for goofusmax_54890

Short answer: not always—but it depends on the product and the specifics.

What is always prohibited

  • Any sexual content involving minors or the sexualization of minors (including grooming, nudity, “age‑inappropriate material” targeting minors). Also prohibited: facilitating “sexual services,” and unauthorized sexualization when impersonating someone. These rules apply platform‑wide. (elevenlabs.io)

Product‑specific restrictions

  • Conversational AI (BYO‑LLM): all sexual content is disallowed, regardless of context. (elevenlabs.io)
  • Music: use by “adult entertainment or pornographic content” industries is prohibited; outputs may not be obscene or unlawful or glorify sexual violence or exploitation of minors. (elevenlabs.io)
  • Voice Library/Live Moderation: when enabled on a shared Professional Voice Clone, requests are checked against prohibited categories. Separately, for Conversational AI, ElevenLabs describes message‑level live moderation that currently blocks sexual‑content‑involving‑minors (SCIM) by default. (help.elevenlabs.io, elevenlabs.io)

How enforcement works

  • ElevenLabs uses automated systems and human review to block or flag content and may take action after generation as well (e.g., warnings, removals, suspensions). (elevenlabs.io)

What this means for “slightly sexual” adult dialogue

  • If it involves consenting adults, does not target minors, and isn’t otherwise unlawful/obscene, the Prohibited Use Policy does not ban it across all services. However, it would still be disallowed in Conversational AI (BYO‑LLM), and could be blocked by live moderation if a shared voice’s owner enabled stricter filters. (elevenlabs.io, help.elevenlabs.io)

If you want us to check a specific generation that triggered a warning, please share the timestamp/voice used and contact Support so they can review the exact event. (help.elevenlabs.io)

Sources:
ElevenLabs Prohibited Use Policy | ElevenLabs
ElevenLabs — Discover our Conversational AI Terms | ElevenLabs
Music Terms | ElevenLabs
What is Live Moderation? – ElevenLabs
An overview of Safety framework for AI voice agents | ElevenLabs
ElevenLabs — Safety | ElevenLabs
ElevenLabs — Discover our Conversational AI Terms | ElevenLabs
What is Live Moderation? – ElevenLabs
How can I contact support? – ElevenLabs

wicked plover
#

To help pinpoint what message was the issue, how long do these automated warnings usually take after the content is generated?