Content Moderation / Knowledge Center / Guide
Trust & Safety changelog
Stay on top of the latest changes affecting your Trust & Safety operations.
This changelog is crowdsourced, feel free to suggest missing items.
Adult contentChild SafetyExtremismHarassmentHateMisinformationPrivacySelf-HarmSubstancesViolence
August 2024 | ||
| TikTok is launching its Sub-Saharan Africa Safety Advisory Council to protect children from threats. Read more | WorldTikTok | ProductChild Safety |
July 2024 | ||
| Meta's Oversight Board is asking the platform to refine its policies around AI-generated explicit images. Read more | WorldMeta | ProductAI |
June 2024 | ||
| Snapchat is rolling out new safety features aimed at protecting teens from unwanted contacts. Read more | WorldSnapchat | ProductChild Safety |
| YouTube is releasing a new feature allowing users to remove AI content that mimics their face or voice. Read more | WorldYouTube | ProductAI |
| LinkedIn is disabling a feature allowing advertisers to target European users based on membership in groups. Read more | EU | ProductPrivacy |
| WeChat is requiring all creators on the platform to disclose whether a published post was generated using AI. Read more | WorldWeChat | ProductAI |
May 2024 | ||
| Twitch is disbanding its Safety Advisory Council and is replacing it with Twitch Ambassadors. Read more | WorldTwitch | Product |
| OpenAI is creating a safety and security committee led by senior executives. Read more | WorldOpenAI | Product |
| TikTok is introducing "Content Credentials", a digital watermark to label AI-created images and videos. Read more or even more | WorldTikTok | ProductAI |
April 2024 | ||
| TikTok is updating its guidelines to restrict weight loss content. Read more | WorldTikTok | ProductSubstances |
| Tinder is adding a new safety feature to the app, so online daters can let friends and families know where they are. Read more | WorldTinder | Product |
| Meta is testing features that blur messages containing nudity to safeguard teens on Instagram. Read more | WorldMeta | ProductChild Safety |
| Meta is announcing an update to its AI labeling policy, expanding its definition of “manipulated media” to go beyond AI-generated videos. Read more or even more | WorldMeta | ProductAI |
March 2024 | ||
| TikTok is unveiling a global Youth Council,aiming at improving younger user safety. Read more | WorldTikTok | ProductChild Safety |
January 2024 | ||
| Substack is implementing a new “report” button in its app, allowing readers to flag posts and publications directly. Read more or even more | WorldSubstack | Product |
| Meta is restricting teens from viewing content that deals with topics like suicide, self-harm, and eating disorders. Read more | WorldMeta | ProductChild SafetySelf-harm |
December 2023 | ||
| Meta is rolling out end-to-end encryption for calls and messages across Facebook and Messenger. Read more | WorldMeta | Product |
| Meta is planning to disable its cross-messaging feature between Facebook and Instagram to comply with the Digital Market Act. Read more or even more | WorldMeta | Product |
November 2023 | ||
| Meta is updating its political advertising policies to cover AI-generated imaged and videos. Read more | WorldMeta | ProductMisinformation |
| TikTok is taking action to remove videos promoting terrorism from the platform. Read more | WorldTikTok | ProductExtremismViolence |
| Meta's Oversight Board is reviewing the app's handling of a video showing an unveiled woman in Iran. Read more | WorldMeta | Product |
| Google is launching Notes, an experimental feature that allows users to share human insights on search results. Read more | WorldGoogle | Product |
| YouTube is launching new features to prevent harmful content exposure for teens. Read more | WorldYouTube | ProductChild Safety |
October 2023 | ||
| X is announcing that posts corrected by Community Notes will now become ineligible for revenue share. Read more | World | ProductMisinformation |
| Meta is declining its Oversight Board's advice from August 2023 to tighten oversight of drug-related posts. Read more | WorldMeta | ProductSubstances |
| Discord is unveiling a system called “Teen Safety Assist” that will warn minors when they get a suspicious message. Read more or even more | WorldDiscord | ProductChild Safety |
| Meta and TikTok are announcing they took action to counter misinformation following the attack by Hamas on Israel. Read more or even more | WorldTikTok | ProductMisinformation |
| Snapchat is facing accusations over its failure to assess privacy risks of its generative AI chatbot "My AI". Read more | WorldSnapchat | ProductPrivacy |
| Meta's Oversight Board is planning to open a case involving political deepfakes. Read more or even more | WorldMeta | ProductMisinformation |
September 2023 | ||
| X is disabling a tool allowing users to report electoral fake news. Read more | World | ProductMisinformation |
| The Mental Health Coalition is launching the Safe Online Standards for Kids' Mental Health (S.O.S) initiative to create a rating system across platforms. Read more | World | ProductChild Safety |
| Meta's Oversight Board is urging the platform to improve the distinction between hate speech and criticism of hate speech. Read more | WorldMeta | ProductHate |
| Snapchat is announcing new safety features such as In-app Warnings to protect teens from potential online risks. Read more | WorldSnapchat | ProductChild Safety |
August 2023 | ||
| ADL is urging Meta’s Oversight Board to take action to improve addressing Holocaust denial and distortion. Read more | WorldMeta | ProductHateMisinformation |
| Meta's Oversight Board is announcing Holocaust Denial as a new case for consideration, inviting people and organizations to submit public comments. Read more | WorldMeta | ProductHateMisinformation |
| Meta is revising its Adult Sexual Exploitation policy, permitting sharing of content involving non-consensual sexual touching if posted with the aim of raising awareness. Read more | WorldMeta | ProductAdult |
| Snap is announcing that its users will soon be able to opt out of content personalization. Read more | WorldSnapchat | Product |
| Instagram's users users will now be able to access features without seeing content that's been ranked by Meta's recommendation algorithms. Read more | WorldMeta | Product |
| X is announcing it will remove a protective feature letting users block other accounts, except for direct messages. Read more | World | Product |
| Google is launching its new Transparency Center on which users can learn more about policies, reporting inappropriate content or appealing a ban. Read more | WorldGoogle | Product |
| YouTube is updating its policies to tackle medical misinformation and remove harmful cancer claims. Read more | WorldYouTube | ProductMisinformation |
| TikTok is announcing that its EU users will be able to switch off its content-selection algorithm to comply with the DSA. Read more | EUTikTok | Product |
| Meta's Oversight Board is urging the company to have stricter rules bannning gender-based violence. Read more | WorldMeta | ProductHate |
July 2023 | ||
| Seven big tech companies are agreeing to commit to new standards in security, trust and safety. Read more | World | Product |
| TikTok is expanding access to its research API to Europe and is launching an ads transparency library to comply with the DSA. Read more | EUTikTok | Product |
| Xbox is adding a reactive voice chat moderation, allowing players to capture and submit 60-second audio clips of inappropriate voice chat messages. Read more | WorldMicrosoft | Product |
| Discord is expanding its policies to address generative artificial intelligence that can create fake content and the sexualization of children. Read more | WorldDiscord | ProductChild Safety |
June 2023 | ||
| TikTok is launching a new feature enabling parents to filter out videos they don't want their children to see. Read more | WorldTikTok | ProductChild Safety |
| TikTok is developing a youth council to build safety tools that are more effective for teenagers. Read more | WorldTikTok | ProductChild Safety |
| Meta is rolling back its measures to curb the spread of Covid misinformation. Read more | WorldMeta | ProductMisinformation |
| Meta is implementing new updates requiring advertisers to designate the beneficiary and the payer for their ads. Read more | WorldMeta | Product |
| The Global Project Against Hate and Extremism (GPAHE) is releasing a database compiling more than 300 hate and far-right extremist symbols. Read more or even more | World | ProductExtremismHate |
| Apple is annoucing that a new feature will soon warn users when receiving unsolicited nudes. Read more | WorldApple | ProductAdult |
| YouTube is announcing it will no longer remove videos with false claims of fraud in the 2020 presidential election. Read more | WorldYouTube | ProductMisinformation |
May 2023 | ||
| Twitter is launching Community Notes for images in posts to address misleading images and emphasize crowdsourced moderation. Read more | WorldTwitter | ProductMisinformation |
| Mozilla is launching a new Mastodon server to test content moderation that is different to other social media platforms, with no focus on free speech or ‘neutrality’. Read more | WorldMozilla | Product |
April 2023 | ||
| TikTok is changing its misinformation policy to ban all climate change denial content on its platform. Read more | WorldTikTok | ProductMisinformation |
| Meta's Oversight Board is asking the social media company to keep its covid misinformation policy and to be more transparent when removing content. Read more | WorldMeta | ProductMisinformation |
| YouTube is updating its guidelines for eating disorder-related content, banning content including extreme calorie counting or purging after eating. Read more | WorldYouTube | ProductSelf-harm |
| Twitter is removing a policy prohibiting the targeted deadnaming or misgendering of transgender people from its moderation guidelines. Read more | WorldTwitter | ProductHate |
| Twitter is planning to add visible labels on tweets that have been identified as potentially violating its policies. Read more | WorldTwitter | Product |
| Snapchat is launching new tools to improve its AI chatbot, including an age filter ensuring it responds according to the user's age. Read more | WorldSnapchat | ProductChild Safety |
| Facebook is planning to roll out new policies making users able to have greater control over "demoted" and fact-checked content. Read more | WorldMeta | Product |
March 2023 | ||
| Twitter is opening some source code to public inspection, including the algorithm used to recommend tweets to users. Read more | WorldTwitter | Product |
| Meta is rolling out a new system to separate ads from harmful or controversial content. Read more | WorldMeta | Product |
| Teleperformance is resuming its full-service content moderation services, including moderation of highly egregious content. Read more | WorldTeleperformance | Product |
| Yubo and the AFNOR Group are forming a working group to create a new safety standard on the prevention of risks and protection of minors on social networks. Read more or even more | WorldAFNORYubo | ProductChild Safety |
| TikTok is unveiling its updated community guidelines that will focus on improving content moderation on the platform and take effect on April 21st. Read more | WorldTikTok | ProductChild SafetyMisinformation |
| Snapchat is publishing guidelines in its Family Center, detailing how content gets algorithmically recommended to minors. Read more | WorldSnapchat | ProductChild Safety |
| Meta is introducing a new way for users to authenticate their account with a missed call. Read more | WorldMeta | ProductPrivacy |
| Meta's oversight board is reviewing the moderation of the Arabic word "shaheed", meaning "martyr" in English, the word associated with most content removals. Read more | WorldMeta | ProductExtremism |
| Whatsapp is agreeing to be more transparent on changes to its terms of service according to the European Commission. Read more or even more | WorldMeta | Product |
| Facebook is revamping its "cross-check" moderation system after facing criticism for applying different review processes for VIP vs regular users. Read more | WorldMeta | Product |
| Twitter is now prohibiting "wishes of harm" in its new violent speech policy, banning users expressing desire for harm. Read more | WorldTwitter | ProductHateViolence |
| TikTok is setting a new default 60-minute daily screen time limit for minors. Read more | WorldTikTok | ProductChild Safety |
February 2023 | ||
| Meta is reforming its penalty system "Facebook Jail", saying users will now be receiving a warning first for most violations. Read more | WorldMeta | Product |
| TikTok is announcing the creation of new European data centers to stay in compliance with EU DSA rules. Read more | EUTikTok | ProductPrivacy |
| Meta's oversight board is announcing it will now review more types of content moderation cases and publish some decisions on an expedited basis. Read more | WorldMeta | Product |
| Meta is rolling out a new version of its ad-matching tool, providing more information about how users activities are feeding ML models. Read more | WorldMeta | ProductPrivacy |
| Google is expanding its misinformation "prebunking" initiative to Germany. Read more | GermanyGoogle | ProductMisinformation |
| Meta is launching new comment moderation tools allowing creators on Facebook to view moderation statistics and manage conversations. Read more | WorldMeta | Product |
| Google is introducing a blur feature helping users avoiding explicit images while using the search engine. Read more | WorldGoogle | ProductAdultViolence |
| TikTok is opening transparency and accountability centers to visitors. Read more | WorldTikTok | Product |
| TikTok is rolling out a revamped account enforcement system, including a new strike system and features dealing with recommendations. Read more | WorldTikTok | Product |
January 2023 | ||
| Twitter is planning to limit permanent suspensions of accounts breaking its rules, adding that any user will be able to appeal an account suspension. Read more | WorldTwitter | Product |
| Meta's oversight board is announcing the removal of the Ukrainian far-right military group Azov Regiment from its list of dangerous individuals and organizations. Read more | WorldMeta | ProductExtremism |
| Meta's oversight board is sharing their decision to redefine Facebook and Instagram's community rules regarding nudity in a less discriminatory way. Read more or even more | WorldMeta | ProductAdult |
| TikTok is announcing creators are now able to restrict their videos to adult viewers. Read more | WorldTikTok | ProductChild Safety |
| Google is developing a free moderation tool for terrorist material identification and removal for smaller platforms. Read more | WorldGoogle | ProductExtremism |
December 2022 | ||
| Meta is launching HMA, a new free tool helping platforms identify and remove violating content. Read more or even more | WorldMeta | ProductExtremism |
| Apple is cancelling its plan to scan photos stored in iCloud to detect CSAM. Read more | WorldApple | ProductChild Safety |
| Meta's oversight board is suggesting that Meta's cross check programme is more commercially driven than commited to human rights. Read more | WorldMeta | Product |
| TikTok and Bumble are joining Meta to stop revenge porn by blocking images from StopNCII.org's bank of hashes. Read more or even more | WorldTikTokBumbleMeta | ProductAdult |
November 2022 | ||
| Naver Z is introducing its Safety Advisory Council providing expertise on their policy and features. Read more or even more | WorldNaver | Product |
| Twitter is announcing its Covid misinformation policy is no longer enforced. Read more or even more | WorldTwitter | ProductMisinformation |
| Teleperformance is announcing it will no longer accept any new highly egregious content moderation work. Read more | WorldTeleperformance | Product |
October 2022 | ||
| TikTok is introducing new and updated features and policies for its LIVE community. Read more | WorldTikTok | ProductChild Safety |
| Twitter is reviewing its policies around permanently banning users. Read more | WorldTwitter | Product |
| Spotify is acquiring Kinzen, a firm specialized in identifying harmul audio content. Read more | WorldSpotify | ProductHateMisinformation |
September 2022 | ||
| Tumblr is creating community labels allowing users to avoid seeing unwanted content. Read more | WorldTumblr | ProductAdultSubstancesViolence |
| Twitter is opening the Twitter Moderation Research Consortium (TMRC) to researchers. Read more | WorldTwitter | Product |
| Instagram is developing a feature protecting users from receiving unsolicited nude photos. Read more | WorldMeta | ProductAdult |
| Facebook is experimenting with asking 250 users to help moderate climate speech. Read more | WorldMeta | ProductMisinformation |
| YouTube is announcing updated content moderation policies to prohibit violent extremist content. Read more | WorldYouTube | ProductExtremism |
| Twitter is expanding its fact-checking feature Birdwatch, allowing users to add additional context to tweets. Read more | WorldTwitter | ProductMisinformation |
August 2022 | ||
| OpenAI is introducing a content moderation endpoint assessing whether the content is sexual, hateful or promoting self-harm. Read more | WorldOpenAI | ProductAdultHateSelf-harm |
| TikTok is said to be training its moderators to detect CSAM using graphic images and videos as a reference guide. Read more | WorldTikTok | ProductChild Safety |
See missing items? You can add them here.
Read more
Towards a fine-grained and contextual approach to Nudity Moderation
Blanket bans on nudity and specifically on bare breasts are coming under increased scrutiny. We hear they clash with cultural expectations and impede right to expression for women, trans and nonbinary people.
AI images: testing human abilities to identify fakes
Results and insights from our AI or not game: how well humans identify AI images, when they get fooled and what we can learn from this.