I’m excited and inspired to be joining a company whose mission is to democratize AI and build the foundation to activate humanity’s potential. Transparency and openness are also core values at
@StabilityAI
, which is critical for building trust.
We are thrilled to announce the addition of renowned trust and safety leader
@ellagirwin
to the Stability AI team as our first SVP of Integrity.
You can learn more about Ella and her role at Stability AI here:
@jack
@elonmusk
@Cernovich
@annecollier
@eirliani
@podesta_lesley
I wish this was false but my experience this year supports this. I fought hard to get funding to replace the people working on this who left in early 2022 and was told no. At one point there were 0 engineers and very few employees working on CSE and still no funding.
@TheRabbitHole84
Our team has been carefully reviewing thousands of suspended accounts for the past month, including accounts noted in this thread. Users that did not engage in threats of harm/ violence, fraud or other illegal activity are being reinstated. Will need ~30 days to finish reviewing.
We have been working hard to develop a balanced approach to free speech and keeping people safe from ongoing, repeated harassment. Over the next few weeks we will make it easier to report targeted harassment as well.
Just kidding folks. 😂 There’s no thread. In all seriousness, I did resign but this has been a once in a lifetime experience and I’m so thankful to have worked with this amazing team of passionate, creative and hardworking people. Will be cheering you all and Twitter as you go!
So one or two people noticed that I left Twitter yesterday. I know there’s been a lot of speculation regarding what happened. Was I fired? Did I quit? Why?? Here’s what really happened: 🧵1/24.
@alx
@PapiTrumpo
@elonmusk
@TwitterSafety
Account was not adhering to our parody policy which is what triggered the suspension however this was an overly strict application of the policy given language and other factors. Team is correcting this action.
@elizableu
Thank you! My team normally only hears what’s going wrong. That’s ok since our job is to fix the things that are broken and protect our users, but it’s awesome to hear the occasional positive feedback, especially because everyone is working so hard!
@benshapiro
@MattWalshBlog
We locked down the account earlier but on a positive note it was great to see the Twitter community come together so quickly to alert us to the issue. Appreciate everyone’s help!
Good reminder for our users on how to appeal an account suspension. Also, looking forward to our new transparency labels for tweets with restricted reach, which start rolling out on Monday. Will take time to roll out for all violations but we are moving fast to get to 100%.
Posting a reminder on how to appeal a suspended account along with a few tips. We are working appeals in less than 48 hours in most cases, but we do require that a user appeals their suspension directly (you can't appeal on someone's behalf). Some tips for a successful appeal…
@stillgray
Correct. We had to automatically sweep our platform and remove >5000 tweets /retweets of this poster. We do not support tweets that incite violence irrespective of who posts them. "Vengeance" does not imply peaceful protest. Organizing or support for peaceful protests is ok
@thetomzone
Correct. For security purposes, the screenshots requested came from me so we could ensure no PII was exposed. We did not give this access to reporters and no, reporters were not accessing user DMs.
Twitter Safety will be posting an update on general amnesty reinstatements this week. As a reminder, we will not reinstate users who engaged in threats of harm or violence, fraud or other illegal activity and we will suspend accounts for this type of activity immediately.
@TheRabbitHole84
Our team has been carefully reviewing thousands of suspended accounts for the past month, including accounts noted in this thread. Users that did not engage in threats of harm/ violence, fraud or other illegal activity are being reinstated. Will need ~30 days to finish reviewing.
In the past 2 weeks, we’ve suspended multiple accounts and/or restricted content, causing confusion for users. Unfortunately, we can’t answer questions or share details about specific users and account actions, but here are some things that can lead to these types of actions:
Thank you
@elizableu
for hosting this Twitter Spaces event and thank you
@andst7
for your partnership and dedication to preventing harm to children globally.
@RealJamesWoods
@elonmusk
We don’t allow any videos or images showing physical or sexual child abuse on our platform even if it is to raise awareness of the crime having occurred. The video was restricted for that reason from yours and other accounts posting it once we became aware of it.
Hate speech impressions (# of times tweet was viewed) continue to decline, despite significant user growth!
@TwitterSafety
will publish data weekly.
Freedom of speech doesn’t mean freedom of reach. Negativity should & will get less reach than positivity.
@videotech_
@elonmusk
Based on our investigation it appears your email account credentials have been compromised. The bad actor was able to access your emails to pass security checks for log in and profile changes. We highly recommend changing your email password and using an authenticator app for 2FA
@WIRED
We do not allow the solicitation of or distribution of hacked/ illegally obtained materials nor do we allow threats to expose or doxx someone through the sharing of private communications/ information. The reporter suspended was doing all of the above based on their tweets.
I fly so often that it has become routine, but sometimes I make a point to watch the planes in front of ours takeoff. Never ceases to amaze me that humans were able to take flight and that we now just mostly sit back and ignore the whole thing like it’s nothing.
@stillgray
@nicksortor
@elonmusk
When a mass shooting occurs, we take immediate action to stop the distribution of any content produced by the perpetrator of the violent attack on Twitter. The restricted tweet included material created by the perpetrator, which violates our policy.
We appreciate the input we have received from our users! Today we developed an approach that will leave tweets up, but will also keep media we don't allow for possible incitement of violence from rendering. We've started applying these changes to recently impacted tweets.
Sharing information surrounding media that we have determined will not be allowed based on our policies. This thread is intended to cover the most common questions being asked.
@DrLoupis
@elonmusk
@Twitter
Your account has not been “destroyed”. It was temporarily (~12 hrs) marked as having sensitive content because you posted an image with nudity and did not label it as containing sensitive content. To avoid this in the future, please see instructions here.
First, we will suspend accounts for threats or incitement of violence. This includes, for example, declaring war on an individual or a group of people based on their race, religion, etc. and asking your followers to join you in this fight.
@nickmartin
@TwitterSafety
Reviewed and found an agent error in handling the report of address doxxing. Fixing this now! We take doxxing seriously. If a user's home address was shared on our platform by someone and we took no action, I would like to review the case. Please DM me more information.
@Glinner
@elonmusk
I realize there's a perception that this tweet was restricted due to mass reporting or employees with a hidden agenda but actually it was just one complaint and a brand new agent that made a mistake when reviewing the complaint. They will be going through more training.
A great quote from Jeff Bezos's 2015 Amazon annual shareholder letter.
"As organizations get larger, there seems to be a tendency to use the heavy-weight Type 1 decision-making process on most decisions, including many Type 2 decisions."
We don’t suspend users for posting reported content unless it is clear the user knew the content was illegal. We do ask for it to be removed before account use is resumed. Soon we’ll display the specific removal reason publicly.
Happy Thanksgiving everyone! Our team has been working really hard removing bad actors, ensuring users are not impacted by hate speech and generally ensuring a safe platform for users. I’m so grateful for everyone on our team and their dedication to Twitter!
@HXMnCK
My team has been working hard to find and remove account level visibility filters from accounts over the past few months. We prefer tweet level actions over account actions. Most have been removed at this stage, however we hope to be done cleaning up the few remaining in May.
We’ve had to suspend accounts, remove documentaries, videos, tweets and photos recently when the proper reporting process was followed. People often disagree with our actions and don’t have full context, but we still have to take action.
@s8n
We do allow satire and we try to reject legal demands for removal of satirical accounts or content. Sometimes we are successful - not always. Not everyone has a sense of humor.
@MaximeBernier
@elonmusk
We show that on 5/24 you tweeted an NSFW image, which caused your account to be correctly labeled as containing sensitive content. This means it won’t be suggested to users who don’t want to see sensitive content, but that is temporary provided no further NSFW content is posted.
These issues are complex and often resolved in court. We don’t determine who in a legal / criminal case is ultimately right or wrong but we will continue to ensure we protect users, comply with local laws and remove content when required.
@zoo_bear
@Delhiite_
@TwitterSupport
The user posted a graphic video depicting an illegal act in violation of our terms of service. We have now reviewed and confirmed that the video was posted with the intent to raise awareness. As long as the video is deleted, the user can resume using the account.
Reports of non-consensual nudity, copyright violations, illegal content, etc. do not get processed by all platforms at exactly the same time, however if properly filed, the content is normally removed from other platforms as well
@AtypicalChad
@elonmusk
Policy is specific about “unreciprocated” exchanges being a criteria. If you are actively engaging and debating a topic with someone you are engaging in a free speech discussion. Just don’t get violent.
Second, we restrict accounts /content when we receive a properly filed report that content was obtained illegally, violates the law or someone’s rights. These are often legal notices and court orders. Improperly filed reports are rejected.
@MrAndyNgo
True. We do need an option to report ban evasion. Trying to strike the balance between providing all reasons to report vs long, complicated reporting flow. We are continuing to work on reporting flow and welcome suggestions on what users think would make it easier / clearer.
@mac78251
We have been digging into this issue for the past few weeks and will be putting some things in place to better detect bad faith reports and reporters faster.
@YourWebmiester
@RealJamesWoods
@elonmusk
Please report these videos/ tweets if you are seeing them. Our team removes many thousands of policy violating, graphic and illegal videos daily but we don’t detect 100% of them on our own.
Cattle dog: Wake up hooman! It’s 6am. I let you sleep in today. Let’s gooooo! I gotz lots of energy!
Dog owner: (half awake takes dogs out for a run. Now wide awake)
Cattle dog: (1 hour later):
@FrankP614
@elonmusk
@TwitterSafety
Wrong message was selected by agent. It should have said that you can't use our blue checkmark logo in your profile picture. Only way to get a blue checkmark is to actually subscribe and be approved.
@MishaFitton
@LindYaccarino
@TopLobsta
The account was not suspended over this tweet. We don't yet consider launching someone into space to be an actual threat. The user was suspended for a different tweet and has appealed. Team is reviewing the appeal.
@sparklingruby
Bad faith reporter = a user reporting other users en masse in order to have their tweets/ account taken down. For example someone submitting 1000 reports per day on every user they don't like even though that user hasn't done anything policy violating.
Sharing information surrounding media that we have determined will not be allowed based on our policies. This thread is intended to cover the most common questions being asked.
@UnWokablePod
The graphic was reported by a high number of users across our platform yesterday and yes, I'm sure the timing of that was due to heightened sensitivity to the language, given the tragic events in Nashville. We always evaluate tweets driving a sudden spike in user complaints.
@AldoButtazzoni
@elonmusk
Team researched and found a bug impacting your account as well as several others. They are fixing it now. Should resolve the issue with your handle not appearing in search results. No other issues identified.
@2VNews
Doing that too. We have already suspended ~100 users in the past month who submitted over 500k bad faith reports this year and will be actively enforcing users who weaponize our reporting system going forward.
Thank you
@jaysullivan
. I’m excited to join as the new product lead for Health and Twitter Service. Looking forward to helping continue and evolve the great work that teams across
@Twitter
have been doing to ensure healthy conversations and a great user experience.
I’m thrilled to welcome
@ellagirwin
as
@Twitter
's new VP of Product for Health and Twitter Service. An inspiring and experienced leader, Ella will play a key role in helping us safeguard the health of the public conversation on Twitter. Welcome to the flock!
@jonesville
@troonytoons
@SukiTrans
@elonmusk
We can't allow videos of an illegal sexual act to be posted on Twitter, even if done to raise awareness. There were also women filmed in a restroom who were not aware they were being recorded. We have since confirmed the identity of the user involved and suspended the account.
@stillgray
@crabcrawler1
Account was suspended after a fairly extensive investigation. We can’t comment in any detail on legal / criminal cases but the account has violated multiple privacy rules repeatedly and across multiple accounts.
@ATX_PATRIOT
@elonmusk
It looks like you have received multiple copyright complaints for posting videos with songs from various artists. I do not see any copyright complaints involving images of your dog as the reason for the complaint.
@MrAndyNgo
@elonmusk
This is actually false. I would recommend checking information like this before posting. Jeff stepped into this role as part of Twitter 2.0.
@stormrobinson
Thx for recognizing the faster response times! I wish I could say this was a win since the team has been working hard making changes, but both suspensions were errors so we still have work to do here. We apologize for the frustration and confusion!
@robbystarbuck
We share the concerns. We are deploying an alternative solution tonight, similar to what we discussed that will improve over time but hopefully strike a better balance surrounding free speech while also ensuring we don’t make it easy to organize/incite violence on our platform.
Is there anything like driving back roads, windows down, favorite music playing to clear your head? Not ready to give up the driving to the machines just yet.
@flickorr
@stillgray
I will review the copyright complaints involved and confirm if suspension action was correct. I like animal videos too but there are copyright laws and ongoing, obvious infringement is not something we can allow. If the account is eligible we will reinstate with a warning tonight
Watching this whole team come together, take on more, link arms across verticals and countries and functions just to get sh*t done - has been absolutely remarkable.
After all this,
#OneTeam
is still alive and well.
New Twitter policy is freedom of speech, but not freedom of reach.
Negative/hate tweets will be max deboosted & demonetized, so no ads or other revenue to Twitter.
You won’t find the tweet unless you specifically seek it out, which is no different from rest of Internet.
@rebeccamkern
@TwitterSafety
Because it was unnecessary to single any specific group out as an example. The sentence “we prohibit behavior and content that harasses, shames, or degrades others.” is sufficient to cover abuse of trans individuals and many other abuse and harassment situations.
@MrAndyNgo
@elonmusk
Yes. Jeff works for me, which is why I can correct this info. He has been at Twitter since 2021 but working in a different role. He recently stepped into his new role as part of Twitter 2.0 and has been integral in helping us drive transparency and fairness.
@EoinLenihan
@TwitterSafety
Doxxing or threats to expose private information should be reported by reporting the tweet or user (second bullet in reporting flow lists privacy violations as included). We did not find any reports of this tweet having been reported previously, but have now taken action.
@Rainmaker1973
Copyright violations can happen intentionally and unintentionally so we evaluate multiple factors when reviewing legal claims. We may have to ask you to remove content if the legal claim is valid but suspensions for DMCA are limited to pirating / egregious illegal behavior.
@Rationalist69
@stillgray
We did not take a "strike" approach with restricting these tweets. We just restricted the media being shared. There is no impact to users for having tweeted it unless it is reposted after removal or was posted with additional calls for violence/ wishes of harm.
@onlinecolloquy
@TheRabbitHole84
@elonmusk
@TrueIndology
We certainly appreciate the passion surrounding specific accounts. There are a lot to review and it will take ~30 days to get through all accounts.
@trueindology
is coming up for review, but we don’t move someone to the front of the line based on number of tweets they get.
@DonnaRiddell
@Rainmaker1973
Account is not at any risk for suspension. I’m not sure there is any specific help needed here from my team but let me know if something is pending / needs more clarification.
@Junglethc420
@TwitterSupport
@TwitterSafety
@elonmusk
We never said we have fully solved the bot problem, but we have invested heavily in reducing bots on our platform and their ability to cause harm. The examples you used are not bot accounts and the source you are using for this is not credible data.