AI Usage at a Glance
Mar 20, 2015
RecommendationPractice documented: Everything a user sees when they open Pinterest is chosen and ranked by AI. The system — called Pinnability — predicts how likely each individual user is to engage with each Pin and orders content accordingly. No human editor curates the home feed. The system has been continuously updated since 2015; the most recent major version, TransActV2, launched in 2025 and personalizes results using more than 16,000 historical user actions per person.
Practice DocumentedView practice →Jun 6, 2018
RecommendationPractice documented: Pinterest built a specialized AI model called PinSage that figures out which Pins are related to each other by analyzing the structure of how billions of users have organized content into boards. If many different users have saved both a recipe for sourdough bread and a specific mixing bowl to their cooking boards, PinSage learns those items are related — even if they look nothing alike visually. PinSage was first described in 2018 and its underlying approach still powers Pinterest's related content recommendations, search, and shopping features today.
Practice DocumentedView practice →Aug 15, 2018
RecommendationNew evidence: PinSage: A new graph convolutional neural network for web-scale recommender systems
Evidence AddedView practice →Feb 26, 2019
ModerationPractice documented: Pinterest uses AI classifiers and URL-level blocking to prevent the spread of health misinformation — including content questioning vaccine safety — and became the first major social media platform to ban climate change misinformation outright in April 2022. Searches for vaccination-related topics are restricted to results from authoritative health sources like the WHO, CDC, and the American Academy of Pediatrics only. These policies have been in place in various forms since 2017.
Practice DocumentedView practice →Jul 22, 2019
ModerationPractice documented: Pinterest automatically detects content that promotes self-harm or suicide and removes it. When a user searches for terms related to emotional distress or self-harm, the platform intercepts the search and shows wellbeing resources — such as breathing exercises and crisis hotline information — instead of or alongside regular results. The crisis search feature, called Compassionate Search, was developed with Stanford Medicine's Brainstorm Lab and crisis organizations, and launched in July 2019.
Practice DocumentedView practice →Jul 23, 2019
ModerationNew evidence: Pinterest launches new search tool to help users' mental health
Evidence AddedView practice →Oct 28, 2020
ModerationNew evidence: Externally polished, internally marginalized: Inside the fight over Pinterest’s health misinformation policy
Evidence AddedView practice →Jan 22, 2021
OtherPractice documented: Pinterest offers augmented reality (AR) try-on tools that let users see what products would look like on themselves or in their homes before buying. Users can try on thousands of lipstick and eyeshadow shades using their phone camera, or see how furniture and home décor items would look in their actual room. The beauty try-on launched around 2020 and the home décor version in January 2022. Pinterest says users are five times more likely to show purchase intent on Pins with try-on enabled.
Practice DocumentedView practice →Mar 5, 2021
ModerationPractice documented: Pinterest uses machine learning to automatically scan billions of images and pieces of text across its platform for content that violates its policies, including adult content, hate speech, self-harm encouragement, medical misinformation, drug promotion, and graphic violence. The system works both in batches — reviewing all existing content daily — and in real time as new content is uploaded. Pinterest says reports of policy-violating content per impression dropped 52% after the system was first deployed in 2019.
Practice DocumentedView practice →May 14, 2021
ModerationPractice documented: Pinterest runs a multi-layered automated system to detect and remove spam content and fake accounts. The system operates in real time — flagging suspicious behavior within seconds of it occurring — as well as in scheduled batch scans of all accounts and content. Pinterest's transparency reports document over one million account deactivations for spam in a single six-month period.
Practice DocumentedView practice →Dec 3, 2021
ModerationPractice documented: Pinterest uses an AI model to automatically screen comments on Idea Pins for unsafe content, spam, and negative or harmful language before those comments become widely visible. The system works in near-real time as comments are posted. Since the system was introduced, Pinterest has reported a 53% decline in policy-violating comment reports.
Practice DocumentedView practice →Jan 31, 2022
OtherNew evidence: Pinterest introduces AR Try On for Home Decor for the ultimate online home shopping experience
Evidence AddedView practice →Apr 6, 2022
ModerationNew evidence: Pinterest bans all climate change misinformation on its platform
Evidence AddedView practice →Nov 4, 2022
RecommendationNew evidence: How Pinterest Leverages Realtime User Actions in Recommendation to Boost Homefeed Engagement Volume
Evidence AddedView practice →Sep 7, 2023
OtherPractice documented: Pinterest uses AI to identify diverse skin tones, body types, and hair patterns across billions of images on its platform, then uses those signals to ensure that search results and recommendations reflect a broader range of human appearances. Users can filter fashion search results by body type ranges, skin tone ranges, and hair patterns. The body type technology launched in September 2023, with public user filters rolling out in March 2024; skin tone ranges have been available since 2018.
Practice DocumentedView practice →Jan 9, 2024
Data AnalysisPractice documented: Pinterest uses a multi-layer AI system to predict which users are most likely to take a specific action after seeing an ad — such as making a purchase or adding something to a cart. This prediction shapes which ads get shown to which users and at what price. The system has been continuously upgraded over several years, shifting from simpler rule-based tools to deep neural networks.
Practice DocumentedView practice →Mar 11, 2024
OtherNew evidence: Pinterest rolls out its ‘body type ranges’ tool to the US
Evidence AddedView practice →Apr 25, 2024
RecommendationPractice documented: Pinterest uses an AI system called OmniSearchSage to power its search results and shopping recommendations. Unlike earlier systems that treated text searches and image searches as separate problems, OmniSearchSage places search queries, Pin images, and product listings into a single shared mathematical space so that a search for "minimalist bedroom" can surface both text-matching articles and visually relevant room photos at the same time. The system serves approximately 300,000 search requests per second..
Practice DocumentedView practice →May 1, 2024
Creative GenPractice documented: Pinterest uses computer vision — the ability of software to understand what's in a photo — to automatically cut objects out of images so users can combine them into collage-style mood boards. Since June 2025, advertisers can use a related tool called Auto-Collages, which automatically turns a product catalog into thousands of shoppable collage ads in minutes. A consumer-facing version called "Styled for you" launched in October 2025, mixing saved fashion Pins into AI-curated outfit ideas.
Practice DocumentedView practice →Jul 10, 2024
Creative GenPractice documented: Pinterest built its own image-generation model, called Pinterest Canvas, that automatically creates lifestyle backgrounds for product photos. Advertisers upload a product image, and the model places it into a contextual scene — like a cozy living room or a kitchen counter — personalized to match styles popular on Pinterest. The tool launched for advertisers in October 2024.
Practice DocumentedView practice →Oct 1, 2024
Creative GenNew evidence: Pinterest rolls out genAI tools for product imagery to advertisers
Evidence AddedView practice →Oct 1, 2024
ProductivityPractice documented: Pinterest launched a suite of AI tools for advertisers called Performance+ in October 2024 that automates most of the steps needed to run an ad campaign. Instead of manually choosing who to target, what to bid, and how to design each ad, advertisers set a goal and the AI handles the rest. Pinterest says the suite cuts campaign setup time in half and improves cost-per-action by more than 10%.
Practice DocumentedView practice →Oct 6, 2024
ProductivityNew evidence: Pinterest Expands Campaign Automation Elements
Evidence AddedView practice →Feb 3, 2025
RecommendationNew evidence: Advancements in Embedding-Based Retrieval at Pinterest Homefeed
Evidence AddedView practice →Apr 4, 2025
Data AnalysisPractice documented: Pinterest uses a fine-tuned AI language model to automatically judge whether search results are relevant to what a user was looking for — a task that was previously done by human raters. The AI can evaluate 150,000 search result pairs in 30 minutes, which would take a team of human reviewers far longer and cost significantly more. This is an internal tool that improves the quality of search results users see.
Practice DocumentedView practice →Apr 30, 2025
ModerationPractice documented: Pinterest introduced a system in April 2025 to automatically detect when an image has been created or significantly altered by AI tools, and labels those images visibly for users browsing the platform. In October 2025, Pinterest added user controls allowing people to reduce how much AI-generated content appears in specific categories of their feed — including beauty, art, fashion, and home décor. The company estimates that approximately 57% of all internet content contains some degree of AI modification.
Practice DocumentedView practice →May 5, 2025
OtherPractice documented: Pinterest Lens is a camera feature in the Pinterest app that lets users point their phone at any object — a piece of furniture, a clothing item, a food dish — and instantly find visually similar Pins and shoppable products. The system recognizes approximately 2.5 billion objects and processes hundreds of millions of visual searches per month. It was first launched in February 2017 and was significantly upgraded in May 2025 with multimodal search capabilities, allowing users to combine a photo with a text description in a single search.
Practice DocumentedView practice →Jun 6, 2025
RecommendationNew evidence: Next-Level Personalization: How 16k+ Lifelong User Actions Supercharge Pinterest’s Recommendations
Evidence AddedView practice →Jun 11, 2025
Creative GenNew evidence: Pinterest tests an AI feature that lets advertisers turn their catalogs into shoppable collages
Evidence AddedView practice →Oct 16, 2025
ModerationNew evidence: Pinterest rolls out new tools to give users more control over GenAI content
Evidence AddedView practice →Oct 21, 2025
Data AnalysisPractice documented: Pinterest uses AI to identify what phase of a project or interest a user is currently in — for example, whether they are just starting to research home renovation ideas or are close to making a purchase decision. This understanding is then used to send more timely notifications and recommendations. The system, described in an October 2025 engineering post, resulted in an 88% higher email click rate and 32% higher push notification open rate compared to previous approaches.
Practice DocumentedView practice →Oct 27, 2025
Creative GenNew evidence: Pinterest experiments with new AI-powered personalized boards
Evidence AddedView practice →Oct 27, 2025
OtherPractice documented: Pinterest is testing AI-powered upgrades to its Boards feature that go beyond simple user-organized collections. In October 2025, Pinterest launched an experiment in the U.S. and Canada where the platform automatically generates outfit ideas from a user's saved fashion Pins ("Styled for you"), creates entirely AI-curated boards based on trending styles ("Boards made for you"), and suggests related products based on what a user has already saved ("Make It Yours").
Practice DocumentedView practice →Oct 31, 2025
Creative GenPractice documented: Pinterest built its own AI model called Navigator-1 to power a conversational shopping tool called the Pinterest Assistant. Users can type, speak, or upload a photo to ask for ideas — for example, describing a room style they want — and the Assistant responds with personalized product recommendations drawn from their saved content. The Assistant launched in beta for U.S. users in October 2025.
Practice DocumentedView practice →Nov 3, 2025
Creative GenNew evidence: How Pinterest’s VLM-Powered AI Assistant ‘Gets’ Your Style
Evidence AddedView practice →Dec 10, 2025
Data AnalysisNew evidence: LLM-Powered Relevance Assessment for Pinterest Search
Evidence AddedView practice →Jan 13, 2026
RecommendationPractice documented: Pinterest uses AI to automatically organize billions of products from retailer catalogs into themed shopping collections — for example, grouping all "90s minimalist sneakers" or "coastal grandmother kitchen items" together — without requiring human merchandisers to do it manually. The system was described in a Pinterest engineering post published in January 2026.
Practice DocumentedView practice →Feb 19, 2026
ModerationNew evidence: Pinterest Is Drowning in a Sea of AI Slop and Auto-Moderation
Evidence AddedView practice →Pinterest maintains an internal program called Responsible AI that covers three areas: building inclusive AI features (the skin tone, body type, and hair pattern tools described in Practice 21), testing its AI systems for unfair bias (using internal tools that measure whether models perform equally well across different groups of users), and developing guidelines for the safe use of generative AI. The program includes red teaming — structured attempts to find flaws or harmful outputs in AI systems before they launch.
The ML fairness component uses internal tooling that segments model performance by user demographic characteristics and visualizes disparities using fairness metrics. Pinterest's engineering teams are required to run fairness assessments on machine learning models before deployment. The generative AI safety component includes evaluation frameworks and mitigation strategies developed in response to risks introduced by generative AI features like Canvas and the Pinterest Assistant. Pinterest participates in the Digital Trust and Safety Partnership (DTSP) GenAI Working Group and co-founded the Inspired Internet Pledge with Boston Children's Hospital focused on child online safety.
Pinterest uses an AI model to automatically screen comments on Idea Pins for unsafe content, spam, and negative or harmful language before those comments become widely visible. The system works in near-real time as comments are posted. Since the system was introduced, Pinterest has reported a 53% decline in policy-violating comment reports.
Pinterest introduced a system in April 2025 to automatically detect when an image has been created or significantly altered by AI tools, and labels those images visibly for users browsing the platform. In October 2025, Pinterest added user controls allowing people to reduce how much AI-generated content appears in specific categories of their feed — including beauty, art, fashion, and home décor. The company estimates that approximately 57% of all internet content contains some degree of AI modification.
Pinterest runs a multi-layered automated system to detect and remove spam content and fake accounts. The system operates in real time — flagging suspicious behavior within seconds of it occurring — as well as in scheduled batch scans of all accounts and content. Pinterest's transparency reports document over one million account deactivations for spam in a single six-month period.
Pinterest uses AI classifiers and URL-level blocking to prevent the spread of health misinformation — including content questioning vaccine safety — and became the first major social media platform to ban climate change misinformation outright in April 2022. Searches for vaccination-related topics are restricted to results from authoritative health sources like the WHO, CDC, and the American Academy of Pediatrics only. These policies have been in place in various forms since 2017.
Pinterest automatically detects content that promotes self-harm or suicide and removes it. When a user searches for terms related to emotional distress or self-harm, the platform intercepts the search and shows wellbeing resources — such as breathing exercises and crisis hotline information — instead of or alongside regular results. The crisis search feature, called Compassionate Search, was developed with Stanford Medicine's Brainstorm Lab and crisis organizations, and launched in July 2019.
Pinterest uses machine learning to automatically scan billions of images and pieces of text across its platform for content that violates its policies, including adult content, hate speech, self-harm encouragement, medical misinformation, drug promotion, and graphic violence. The system works both in batches — reviewing all existing content daily — and in real time as new content is uploaded. Pinterest says reports of policy-violating content per impression dropped 52% after the system was first deployed in 2019.
Pinterest maintains an internal program called Responsible AI that covers three areas: building inclusive AI features (the skin tone, body type, and hair pattern tools described in Practice 21), testing its AI systems for unfair bias (using internal tools that measure whether models perform equally well across different groups of users), and developing guidelines for the safe use of generative AI. The program includes red teaming — structured attempts to find flaws or harmful outputs in AI systems before they launch.
Pinterest is testing AI-powered upgrades to its Boards feature that go beyond simple user-organized collections. In October 2025, Pinterest launched an experiment in the U.S. and Canada where the platform automatically generates outfit ideas from a user's saved fashion Pins ("Styled for you"), creates entirely AI-curated boards based on trending styles ("Boards made for you"), and suggests related products based on what a user has already saved ("Make It Yours").
Pinterest uses AI to identify diverse skin tones, body types, and hair patterns across billions of images on its platform, then uses those signals to ensure that search results and recommendations reflect a broader range of human appearances. Users can filter fashion search results by body type ranges, skin tone ranges, and hair patterns. The body type technology launched in September 2023, with public user filters rolling out in March 2024; skin tone ranges have been available since 2018.
Pinterest offers augmented reality (AR) try-on tools that let users see what products would look like on themselves or in their homes before buying. Users can try on thousands of lipstick and eyeshadow shades using their phone camera, or see how furniture and home décor items would look in their actual room. The beauty try-on launched around 2020 and the home décor version in January 2022. Pinterest says users are five times more likely to show purchase intent on Pins with try-on enabled.
Pinterest Lens is a camera feature in the Pinterest app that lets users point their phone at any object — a piece of furniture, a clothing item, a food dish — and instantly find visually similar Pins and shoppable products. The system recognizes approximately 2.5 billion objects and processes hundreds of millions of visual searches per month. It was first launched in February 2017 and was significantly upgraded in May 2025 with multimodal search capabilities, allowing users to combine a photo with a text description in a single search.
Pinterest uses AI to automatically organize billions of products from retailer catalogs into themed shopping collections — for example, grouping all "90s minimalist sneakers" or "coastal grandmother kitchen items" together — without requiring human merchandisers to do it manually. The system was described in a Pinterest engineering post published in January 2026.
Pinterest uses an AI system called OmniSearchSage to power its search results and shopping recommendations. Unlike earlier systems that treated text searches and image searches as separate problems, OmniSearchSage places search queries, Pin images, and product listings into a single shared mathematical space so that a search for "minimalist bedroom" can surface both text-matching articles and visually relevant room photos at the same time. The system serves approximately 300,000 search requests per second..
Pinterest built a specialized AI model called PinSage that figures out which Pins are related to each other by analyzing the structure of how billions of users have organized content into boards. If many different users have saved both a recipe for sourdough bread and a specific mixing bowl to their cooking boards, PinSage learns those items are related — even if they look nothing alike visually. PinSage was first described in 2018 and its underlying approach still powers Pinterest's related content recommendations, search, and shopping features today.
Everything a user sees when they open Pinterest is chosen and ranked by AI. The system — called Pinnability — predicts how likely each individual user is to engage with each Pin and orders content accordingly. No human editor curates the home feed. The system has been continuously updated since 2015; the most recent major version, TransActV2, launched in 2025 and personalizes results using more than 16,000 historical user actions per person.
Pinterest uses AI to identify what phase of a project or interest a user is currently in — for example, whether they are just starting to research home renovation ideas or are close to making a purchase decision. This understanding is then used to send more timely notifications and recommendations. The system, described in an October 2025 engineering post, resulted in an 88% higher email click rate and 32% higher push notification open rate compared to previous approaches.
Pinterest uses a multi-layer AI system to predict which users are most likely to take a specific action after seeing an ad — such as making a purchase or adding something to a cart. This prediction shapes which ads get shown to which users and at what price. The system has been continuously upgraded over several years, shifting from simpler rule-based tools to deep neural networks.
Pinterest uses a fine-tuned AI language model to automatically judge whether search results are relevant to what a user was looking for — a task that was previously done by human raters. The AI can evaluate 150,000 search result pairs in 30 minutes, which would take a team of human reviewers far longer and cost significantly more. This is an internal tool that improves the quality of search results users see.
Pinterest built its own AI model called Navigator-1 to power a conversational shopping tool called the Pinterest Assistant. Users can type, speak, or upload a photo to ask for ideas — for example, describing a room style they want — and the Assistant responds with personalized product recommendations drawn from their saved content. The Assistant launched in beta for U.S. users in October 2025.
Pinterest uses computer vision — the ability of software to understand what's in a photo — to automatically cut objects out of images so users can combine them into collage-style mood boards. Since June 2025, advertisers can use a related tool called Auto-Collages, which automatically turns a product catalog into thousands of shoppable collage ads in minutes. A consumer-facing version called "Styled for you" launched in October 2025, mixing saved fashion Pins into AI-curated outfit ideas.
Pinterest built its own image-generation model, called Pinterest Canvas, that automatically creates lifestyle backgrounds for product photos. Advertisers upload a product image, and the model places it into a contextual scene — like a cozy living room or a kitchen counter — personalized to match styles popular on Pinterest. The tool launched for advertisers in October 2024.
Have evidence about Pinterest's AI practices? Submit a report.
Submit a report →AI Trace is free and nonprofit. Support our work
Pinterest is testing AI-powered upgrades to its Boards feature that go beyond simple user-organized collections. In October 2025, Pinterest launched an experiment in the U.S. and Canada where the platform automatically generates outfit ideas from a user's saved fashion Pins ("Styled for you"), creates entirely AI-curated boards based on trending styles ("Boards made for you"), and suggests related products based on what a user has already saved ("Make It Yours").
The features use Pinterest's existing recommendation and generative AI infrastructure — including the home feed ranking models and Canvas image generation — applied to the Board surface rather than the main feed. "Styled for you" is a collage-format display that mixes and matches items a user has already saved into outfit combinations they may not have considered. "Boards made for you" are fully machine learning-generated collections with no user curation required. Pinterest is testing these features selectively and has not disclosed the full technical architecture behind each component.
Pinterest uses AI to identify diverse skin tones, body types, and hair patterns across billions of images on its platform, then uses those signals to ensure that search results and recommendations reflect a broader range of human appearances. Users can filter fashion search results by body type ranges, skin tone ranges, and hair patterns. The body type technology launched in September 2023, with public user filters rolling out in March 2024; skin tone ranges have been available since 2018.
The body type system uses computer vision to classify shape, size, and form across more than 5 billion images — a task Pinterest describes as a first for the industry at this scale. These classifications feed into a diversity-aware ranking layer that adjusts search result ordering to prevent any single body type from dominating results. Pinterest reports a 454% improvement in body type representation on women's fashion searches, and users who applied body type range filters had a 66% higher engagement rate per session. The skin tone detection system uses the Monk Skin Tone Scale (a 10-point scale developed by Harvard sociologist Ellis Monk) to classify images, which users can then filter by. Hair pattern search allows users to filter by curl type. The systems were developed in consultation with external advocates including NAAFA and model Tess Holliday.
