Miju Labs

← All buyers · OpenAI

A

Adele Li

Product Lead, ChatGPT Images · OpenAI

medium confidenceTier A · Talk to firstFrontier labsimage · UI

Product lead for ChatGPT Images — credited as Product Lead on GPT-Image-1.5 (December 2025) and the public face of the ChatGPT Images 2.0 upgrade shipped in September 2026. She decides what the image surface is for, which in practice means she defines what a 'good' output is at the level users experience it: infographics, slides, maps, manga, multilingual text.

Background

Li is a product manager at OpenAI attached to the image generation line. Her visible arc tracks the productisation of image models: from 4o native image generation as a chat capability, to GPT-Image-1.5 as a faster and more editable model, to ChatGPT Images 2.0 in September 2026 with a full editor and output types that are closer to design deliverables than to pictures — infographics, slide layouts, maps, dense multilingual typography.

That last step matters for a creative-data vendor. Once the product promises slides and infographics, the quality bar stops being photorealism and becomes layout, hierarchy, typographic craft and information design — judgements that professional designers make and that no automatic metric captures.

What they run now

The ChatGPT Images product: the editor, the output formats, and the quality bar users judge. ChatGPT Images 2.0 (announced early September 2026, covered first by Axios) added a substantially upgraded image editor and capability across infographics, slides, maps and manga.

What they buy

Inference, but with a sharp edge. A product lead shipping infographics, slides and maps needs evaluation sets that look like real design briefs and graders who can say whether the result is publishable. Plausible objects: briefed task sets (brand guidelines plus a request, graded on adherence and craft), designer preference pairs on layout and typography, failure taxonomies for dense-text rendering, and localisation judgements for multilingual typography. She is likely to be the person who feels the absence of these most acutely, because she is the one reading user complaints.

Career

  1. presentProduct (ChatGPT Images), OpenAI — Product Lead credit on GPT-Image-1.5, Dec 2025

On the record

Image generation's moment

She discussed the current wave in image generation as a creative renaissance on OpenAI's own podcast. source

Interviews & talks

Why it matters here

High, and probably the most reachable high-value OpenAI contact in this batch. The move into slides and infographics is the exact overlap between OpenAI's roadmap and Julian's thesis: design judgement, not photo realism. She is senior enough to sponsor an eval-set purchase and junior enough to answer a well-aimed message.

How to reach

Product-shaped, evidence-first. Speculation: take the ChatGPT Images 2.0 infographic and slide output, have a panel of working designers grade a hundred of them against a published rubric, and send her the failure taxonomy — the three things designers consistently mark down. Offer the graded set as a standing eval. Do not open with the roster of designers.

What we could not establish
  • X handle appears to be @adele__li but was not confirmed against a bio naming OpenAI; left null.
  • Exact title and tenure not evidenced beyond the release credit.
  • Education and prior employers unknown.
  • Date of the OpenAI Podcast episode not confirmed.

Why they are on the map

'Product Lead' in ChatGPT Images credits (Dec 2025); Images 2.0/2.5 add templates, posters, sketch editing (design-adjacent) source — evidence dated 2025-12

Decides whether the problem is funded.

How OpenAI buys data

Has a dedicated Human Data team that 'turns human feedback into reliable signals for training and evaluation' via bespoke campaigns with 'external vendors and AI trainers' and has open PM/TPM/EM/TLM roles on it (2025-26 job posts). Ended its Scale AI relationship after Meta's stake (Jun 2025); reported to be a major revenue source for Mercor and to use Surge and Handshake, while building in-house data capacity. GDPval (Oct 2025) and HealthBench (May 2025, 262 physicians) show it buys large expert-graded eval sets. Image models ship with a named 'Data & Evaluation' team (Dec 2025 credits).

A central Human Data team (PMs, TPMs, engineers) works with all research teams and vendors. Glaese, a former Human Data lead, is now VP Research & Safety (Jul 2026). Image generation sits under Multimodal (Dhariwal), with its own data leads and an 18-person Data & Evaluation group, plus Model Behavior. Sora was shut down in Mar 2026 and Peebles left in Apr 2026. Image work was folded into GPT Image 2 / 2.5.

Known vendors: mercor, surge-ai, handshake-ai, scale-ai (ended Jun 2025)

Sources

  1. The new ChatGPT Images is here — credits
  2. Adele Li, LinkedIn
  3. OpenAI gives ChatGPT Images a major upgrade
  4. ChatGPT Images 2.0 coverage