AI Transparency

Where AI is used in CoreClips AI

• Transcription. When you upload a video, speech-to-text AI converts the audio into a written transcript, including speaker attribution (identifying which participant said what).
• Summarization. AI models generate structured summaries, takeaways, and insights from the transcript of your video.
• Highlight generation. AI identifies and extracts candidate highlight segments from your video based on the content of the transcript and your configuration (such as target duration or vertical/topic).
• Export formatting. AI-assisted formatting is used when preparing outputs for export (such as structured PDF or presentation exports).

We use a combination of Critso-operated processing logic and third-party AI providers to perform these tasks. The specific third-party providers currently in use are listed in our Subprocessor List; we reference them only generically here because the specific providers may change over time as we evaluate quality, cost, and reliability, and the list is kept current in that document rather than duplicated here.

Outputs can contain inaccuracies

AI-generated transcripts, summaries, and highlights are produced by automated models and can be wrong. Depending on audio quality, accents, overlapping speech, background noise, technical or specialized vocabulary, or the inherent limitations of the underlying models, you may see:

• Mistranscribed or missing words or phrases;
• Incorrect attribution of speech to the wrong speaker;
• Summaries that omit important context, or that state something with more confidence than the source material supports;
• Highlight clips that miss the most relevant moment, or that cut a clip in a way that changes its meaning relative to the full recording.

AI models do not "understand" content the way a person does — they produce statistically likely outputs based on patterns learned from training data, which is not the same as verified factual accuracy.

Human review is recommended

Because of the limitations above, we recommend that a person review AI-generated outputs before you rely on, publish, share, or act on them — particularly in contexts where accuracy matters most, including (without limitation) legal proceedings, medical documentation, compliance records, financial disclosures, or any use where an error could cause harm or have legal consequence. Treat CoreClips AI outputs as a strong first draft that accelerates your work, not as a final, verified record.

You are responsible for how you use AI-generated outputs

As set out in our Terms of Service (Sections 4.4 and 5.1), you are solely responsible for reviewing, verifying, and deciding how to use any Output generated by the Service, including whether it is fit for a particular purpose, and for obtaining any additional review (such as professional or legal review) that your specific use case requires. Critso does not warrant the accuracy, completeness, or fitness for a particular purpose of any AI-generated Output.

No training on your content

As described in our Privacy Policy, Critso does not use your uploaded content to train its own AI models, and does not permit the third-party AI providers we use to train their models on your content.

Questions

Questions about how AI is used in CoreClips AI can be sent to [PLACEHOLDER: privacy@critso.com] or [PLACEHOLDER: legal@critso.com].