Skip to content
This is NOT an official site of the City of New York. Click here for official AI reports.

NYC Office of Technology and Innovation

NYC - DTPR AI Register

AI-Assisted Video Clip Selection for Social Media

What it is for

Marketing & Personalization
  • Used to select and surface the most engaging moments from DYCD podcast recordings for posting on DYCD social media platforms, optimizing for predicted audience engagement (comments, shares, likes).
Run by
Department of Youth and Community Development (DYCD)
Where
No fixed location
Kept
Retained Not specified
Shared with
Accountable organization, Vendor
Your copy
You cannot see the data it holds about you. What you can do

What it is for

OpusClip is an AI platform used by the NYC Department of Youth and Community Development to analyze podcast recordings and identify the most engaging short clips for posting on social media. The system ranks moments in uploaded video by predicted audience engagement — likelihood to generate comments, shares, and likes. It is used biweekly on publicly intended content, and the output clips are selected by DYCD staff before publication. Training data is proprietary to the vendor and not disclosed.

What it collects and what happens to it

Data taken in

Generated content
Anonymized data
  • Finalized podcast video recordings intended for public viewing, uploaded to the OpusClip platform for analysis. The content is publicly intended and does not contain private personal data.

Processing

Recommendation & Ranking
  • Analyzes uploaded video in the context of current social and marketing trends to detect the most compelling, shareable, and emotionally impactful moments. Outputs ranked timestamps and clip segments scored for predicted engagement.

What it does

Deciding (Analytical AI)
Human decides
  • The system ranks and scores video segments by predicted engagement. DYCD staff review the ranked clips and decide which to export and publish — the AI is advisory, not determinative.

Outputs

A recommendation or prediction
Anonymized data
  • Timestamps and exportable clip segments from the original video, ranked by predicted likelihood to generate audience engagement (comments, shares, likes) when posted to social media. Staff select which clips to use.

Run by

Department of Youth and Community Development (DYCD)
  • DYCD Communications uses OpusClip biweekly to identify the most engaging moments in podcast recordings before posting clips to DYCD social media platforms.

NYC Algorithmic Tools Compliance Report — OpusClip

Built by

OpusClip
  • OpusClip develops and maintains the off-the-shelf web platform, including the AI models and proprietary training data that power the video analysis and clip-ranking capabilities.

NYC Algorithmic Tools Compliance Report — OpusClip

Kept for

Retained Not specified
  • The register does not specify how long uploaded video or generated clip rankings are retained by OpusClip or DYCD. The platform updates regularly as a web service.
  • Duration: Not specified

Shared with

Not available to me
  • The AI's internal scoring and ranking data are not made available to the public or to individuals depicted in or associated with the podcast content. Training data is proprietary to OpusClip and not disclosed.
Available to the accountable organization
  • DYCD Communications staff have access to the ranked clip outputs and timestamps within the OpusClip web platform, which they use to select content for publication.
Available to vendor
  • OpusClip, as the platform provider, has access to uploaded video content and usage data as part of operating the web service. The extent of vendor data access beyond processing is not specified in the register.

Stored

Stored on 3rd Party Cloud
  • Video is uploaded to OpusClip's web platform, implying storage on the vendor's cloud infrastructure. The specific cloud provider and data residency jurisdiction are not disclosed in the register.
  • Duration: Not specified
How to read the colours

Can it identify you?

Anonymized data
Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
Pseudonymous data
Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
Identifiable data
The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.

Who completes the loop?

Human decides
This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
Human executes
This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
Autonomous
This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.

Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.

What you can do

Ask about this system

Questions go to the NYC Office of Technology and Innovation, not the vendor.

Your rights

  • Right to Algorithmic TransparencyThe register discloses that OpusClip analyzes video for engagement potential using AI and proprietary training data. However, the detailed logic of the ranking model is not publicly available. Members of the public can consult the NYC Algorithmic Tools Compliance Report at https://data.cityofnewyork.us/City-Government/Algorithmic-Tools-Compliance-Report/jaw4-yuem for the disclosed information.

Risks and safeguards

  • Societal & cultural harmThe system optimizes for predicted social media engagement metrics, which may favour sensationalist or emotionally manipulative content over substantively informative content, shaping how DYCD communicates with the public.Safeguard: DYCD staff make the final selection of clips before publication, maintaining editorial oversight. The system is applied only to content already intended for public viewing.