Last updated: October 10, 2026
Does X allow AI-generated content?
Yes. X does not ban AI-written posts. The rules that apply are the automation rules: no spam, no duplicate posts across accounts, no unsolicited automated replies, and prior written approval from X for any AI reply bot. Separately, AI media that deceptively fakes a real person or event and can cause serious harm is not allowed.
So does X allow AI generated content? Nothing in X's rules asks how a post was written. What the rules care about is what the account does with it: how often, to whom, on how many accounts, and whether the media misleads. Below are the exact lines from X's Help Center, so you can check them yourself rather than take a blog's word for it.
What the automation rules say
X's automation rules are marked "Updated April 2026". They never use the phrase AI-generated content for ordinary posts. They define automation as "accounts or apps that take repeated actions without a person actively performing them", and an AI agent that posts for you is exactly that. The lines that decide whether AI posting is fine:
"Provided you comply with all other rules, you may post automated posts for entertainment, informational, or novelty purposes."
"You may not post duplicative or substantially similar posts on one account or over multiple accounts you operate."
"Automating these actions to reach many users on an unsolicited basis is an abuse of the feature, and is not permitted. For example, sending automated replies to posts based on keyword searches alone is not permitted."
The rules also cover automated posts about trending topics (not allowed), misleading links (not allowed), and automated likes (not allowed). None of these depend on whether a person or a model wrote the words.
Responsibility stays with you
The automation rules put it plainly for users: "You are ultimately responsible for the actions taken with your account, or by applications associated with your account." If an AI tool posts something that breaks the rules, the account is what gets filtered or suspended.The one place AI is named: reply bots
The automation rules have a section titled "AI-Powered Automated Replies". It is the only part that singles out AI, and it is about replies, not posts:
"Provided you comply with all other rules, you may leverage artificial intelligence (AI) technologies to create automated reply bots that generate dynamic, context-aware responses ... However, to safeguard user experience, prevent potential misuse, and ensure alignment with our rules, the deployment or operation of any AI reply bot requires prior written and explicit approval from X."
On top of that policy, the API itself changed. Since February 23, 2026, a reply created through the self-serve X API is only accepted when the original author summoned your account by mentioning or quoting it. Enterprise access is not affected. So an AI that writes your posts is fine, while an AI that goes looking for strangers to reply to needs X's approval and, on self-serve access, cannot do it at all. The background on how that played out is in did X suspend my bot in the purge.
AI images and video: the misleading media rule
X's synthetic and manipulated media rule now sits inside its authenticity policy (dated April 2025), under the heading "Synthetic and Manipulated Media":
"You may not share inauthentic media, including, manipulated, or out-of-context media that may result in widespread confusion on public issues, impact public safety, or cause serious harm."
Among the examples of manipulated media it lists "media depicting a real person that has been fabricated or simulated, especially through use of algorithms or broader artificial intelligence." Two things narrow this. The media has to be "shared as authentic", and it has to carry the risk of confusion or harm. The same policy then says what is allowed:
"Sharing manipulated or out-of-context media in non-deceptive ways."
An AI illustration for a blog post is not what this policy is about. A realistic AI video of a politician saying something they never said is. If an image could be mistaken for a real photo of a real person or event, say in the post that it is AI-generated. X also says that where it cannot reliably tell whether media is misleading, "we may not take action".
The same policy covers fake personas, listing "stock, stolen or AI-generated profile photos" when used "for the purpose of deceiving others".
Allowed and not allowed, at a glance
Every line below comes from the automation rules or the authenticity policy. None of them turns on whether a human or a model wrote the text.
AI-written original posts
Automated posts are allowed "for entertainment, informational, or novelty purposes" if they follow the other rules.
AI posts based on outside data
Posts built from "sources of outside information" such as an RSS feed or weather data, if you are authorised to publish it.
AI replies to people who opted in
Allowed when the user asked to hear from you, can opt out, and gets one reply per interaction. An AI reply bot also needs X's written approval.
AI images shared honestly
The authenticity policy allows "sharing manipulated or out-of-context media in non-deceptive ways".
The same AI post on several accounts
"You may not post duplicative or substantially similar posts on one account or over multiple accounts you operate."
AI replies found by keyword search
"Sending automated replies to posts based on keyword searches alone is not permitted."
Auto-posting about trends
"You may not automatically post about trending topics on X."
A browser agent driving x.com
Scripting the X website is "non-API-based" automation and "may result in the permanent suspension of your account".
Is there an AI-generated content label on X?
Not for individual text posts. What X has is an automated account label, which is about who runs the account, not how a post was written. X's help page describes it this way: "When an account displays the 'automated' account label you know the account is not human-run." It connects the bot to the human-run account that manages it, and X says the label may also appear on that account's posts.
If you write your own posts with AI help and publish them yourself, you are a person using a tool, and the label is not meant for you. If you run a bot account that posts with no person involved, it is. The label is covered in the X automation rules.
If an AI agent posts for you through the API
Agents that publish through the X API are the case the automation rules were written for. The practical checklist, from the rules above:
- Post through the API, not the website. A browser agent clicking around x.com is the "non-API-based" automation the rules warn can lead to permanent suspension.
- One voice per account. If an agent runs several accounts, they cannot post the same or substantially similar text.
- Original posts, not engagement. No automated likes, no keyword-triggered replies, no auto-posting on trends.
- Consent for actions on someone else's account. The rules say OAuth access alone "does not by itself constitute sufficient consent". The user has to be told what will be automated and agree to it.
- A person reads it before it ships. Not a rule, but the cheapest way to catch a model that repeats itself or invents a fact.
The longer version, including API pricing and which endpoints still work, is in X API rules for AI agents in 2026 and the X automation rules guide. Before an AI draft goes out, the AI slop score checks it for the generic phrasing readers spot quickly.
How OpenTweet fits
OpenTweet publishes through the official X API using an OAuth connection you approve. An agent can create posts through the OpenTweet API or MCP server without a scheduled date, which saves them as drafts, and nothing goes out until it is scheduled or published. OpenTweet does not offer automated likes, follows or reply bots.7-day free trial. Cancel anytime.
Frequently asked questions
Does X allow AI-generated content?
Yes. X does not ban AI-written posts. Its automation rules allow automated posts "for entertainment, informational, or novelty purposes" as long as they are not spam or duplicates. The limits are on behaviour: AI reply bots need prior written approval from X, and misleading synthetic media that can cause serious harm is not allowed.
Do I have to label AI-generated posts on X?
X's published rules do not require a per-post AI disclosure for text. The separate "Automated" label is an account label for bots (X says it may also appear on that account's posts), not a disclosure for AI-written text. The authenticity policy targets media that is shared as authentic but deceptively fabricated, so the safe practice for realistic AI images or video of real people or events is to say in the post that they are AI-generated.
Can I run an AI reply bot on X?
Only with X's permission. The automation rules (updated April 2026) say the deployment or operation of any AI reply bot "requires prior written and explicit approval from X". Separately, since February 23, 2026, self-serve API replies are only accepted when the original author mentioned or quoted your account first.
Can an AI agent post to X for me?
Yes, through the X API. Automated original posts are allowed. The agent should post through an authorised API app rather than by scripting the X website, which the automation rules say "may result in the permanent suspension of your account". It also must not post the same or substantially similar text across several accounts.
Are AI-generated images allowed on X?
Yes, unless they deceive. X's authenticity policy prohibits media "depicting a real person that has been fabricated or simulated, especially through use of algorithms or broader artificial intelligence" when it is shared as authentic and can cause widespread confusion on public issues, affect public safety or cause serious harm. It explicitly allows sharing manipulated media in non-deceptive ways.
Can I use an AI-generated profile photo on X?
Only if you are not using it to deceive. The authenticity policy lists "stock, stolen or AI-generated profile photos" as an example of a fake persona when they are used "for the purpose of deceiving others". An AI avatar on an account that is honest about who runs it is a different case.
Does X reduce the reach of AI-generated posts?
X has not published a rule that downranks text because it was written by AI. The enforcement options X lists, including restricting reach, are tied to policy violations such as spam and misleading media. Generic, repetitive posts tend to perform poorly regardless of who wrote them.
Sources: X Help Center automation rules (marked Updated April 2026), authenticity policy (April 2025) and automated account labels page on help.x.com, as of October 10, 2026 (read from archived copies because help.x.com blocks automated fetches). February 23, 2026 reply change: X developer changelog.
Related reading
The automation rules, the label, and what agents can still do on X.
X automation rules
The full rulebook for accounts and AI agents that post.
Twitter automation rules 2026
What is allowed, what is banned, and the gray areas.
X API rules for AI agents
What an agent can and cannot do on X in 2026.
Did X suspend my bot?
The April 2026 reply bot wave and what to do after it.
AI slop score
Check an AI draft for generic phrasing before it posts.
Let AI draft. You decide what posts.
OpenTweet posts through the official X API, with drafts you can review first. Original posts, threads and scheduling, with no automated likes, follows or reply bots.
7-day free trial. Cancel anytime.