Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteMicrosoft announced its acquisition of Two Hat, an online-community content-moderation company, on October 29, 2021. Microsoft did not disclose the purchase price. Two Hat’s technology was already being used in Xbox, Minecraft and MSN, so the transaction expanded an existing partnership rather than creating a new one.
As of 2026, Microsoft presents the relevant technology through its Community Sift community-safety offering. That product identity is distinct from Azure AI Content Safety, Microsoft’s separate developer service for content-classification APIs.
The acquisition at a glance
| Item | What is established |
|---|---|
| Buyer | Microsoft |
| Target | Two Hat, a content-moderation and online-safety company |
| Announcement | October 29, 2021 |
| Purchase price | Not disclosed by Microsoft |
| Stated purpose | Combine Two Hat’s moderation technology and research with Microsoft teams, products and cloud infrastructure |
| Products named in the announcement | Xbox, Minecraft and MSN |
Microsoft’s announcement said the acquisition was intended to improve online safety and digital wellness, make gaming communities more welcoming, extend moderation across Microsoft consumer services and continue serving Two Hat customers and third-party partners. Microsoft also lists Two Hat in its 2021 acquisition history.
What Two Hat built
Two Hat developed proactive moderation systems that could analyze user-generated material before it reached other users. Depending on configuration, the systems could process text messages, usernames, images, videos, reported content and other interaction types.
Detection and classification
The technology used machine learning, linguistic analysis, configurable policy controls and escalation workflows to identify risks such as cyberbullying, harassment, hate speech, violent threats, child exploitation and other violations of a platform’s rules.
Classification is not the same as enforcement
A model can assign a risk score, flag content or route a case to a reviewer. The platform operator still decides whether to allow, warn, blur, block, remove, suspend or escalate that material. Microsoft makes this distinction explicit for Azure AI Content Safety: its APIs return classification information, while customers implement their own enforcement policies.
Two Hat’s pre-acquisition role in Microsoft products
Microsoft said it had worked with Two Hat for several years before buying the company. Its announcement specifically identified Xbox, Minecraft and MSN as services using Two Hat’s proactive moderation technology.
Rank #2
- Keep track of everything from attendance to test scores
- Spiral bound
- Measures 8-1/2" x 11"
That history matters. Microsoft was acquiring a system already tested in both gaming and non-gaming environments, along with expertise in policy design and moderation operations. The deal was therefore an extension of an established relationship, not a speculative purchase of an unproven product.
Recommended Free Tools
Why Microsoft wanted the company
Gaming communities need high-volume screening
Xbox and Minecraft generate continuous streams of messages and user-created material. Automated detection can screen that volume quickly, while human moderators concentrate on ambiguous or severe cases. The objective is to improve response time and consistency, not to make online abuse disappear.
One capability could serve several Microsoft businesses
The announcement connected the acquisition to gaming as well as broader Microsoft consumer services. MSN’s inclusion showed that the opportunity extended beyond game chat and could support different community rules and risk thresholds.
Microsoft could offer the technology to outside customers
Microsoft also described greater opportunity for Two Hat’s existing clients and third-party partners. That points to a business-to-business strategy in addition to internal use: Microsoft could combine the acquired expertise with its cloud, sales and support infrastructure.
Moderators remain part of the system
Modern moderation products are generally designed to prioritize work for people, not remove every human decision. Microsoft’s current Community Sift materials emphasize escalation of difficult cases and configurable policy guides. In one enterprise case study, Microsoft reports an 88% reduction in moderator workload and a 96% reduction in reported toxic incidents. Those are vendor-reported results from a specific customer example, not independent industry benchmarks.
What happened to Two Hat after the acquisition?
Two Hat is no longer best understood as a standalone product company selling under its original identity. Microsoft now markets the associated community-safety capability as Community Sift. Microsoft’s gaming materials describe it as a platform that can classify, filter and escalate user-generated text, usernames, images, videos and reported content, with voice-related capabilities also referenced on its gaming AI and machine-learning page.
An Xbox transparency report also identifies Community Sift as part of Microsoft’s safety technology suite: Xbox transparency report (PDF). Public sources do not establish that every original Two Hat contract, product name or customer relationship continued unchanged, so those details should not be assumed.
Community Sift and Azure AI Content Safety are different products
| Community Sift | Azure AI Content Safety | |
|---|---|---|
| Primary positioning | Managed community-safety and moderation workflow | Developer APIs and tools for content classification |
| Typical buyer | Game, social or community operator needing policy, escalation and moderation support | Developer or enterprise building moderation into its own application |
| Content described by Microsoft | Text, usernames, images, videos, reported content and voice-related interactions | Text, images and other capabilities that vary by API and feature |
| Enforcement model | Filtering, policy configuration and escalation support | Returns classifications; the customer builds enforcement |
| Pricing signal | Contact Microsoft; public page does not show concrete prices | Microsoft documents F0 and S0 Azure tiers; the migration documentation lists 5,000 free transactions per month for F0 |
Do not describe Two Hat as having simply become Azure AI Content Safety. Microsoft documents Azure AI Content Safety separately, while Community Sift is the current product most directly associated with the Two Hat technology lineage. Microsoft also says the older Azure Content Moderator is being retired in favor of Azure AI Content Safety; that legacy service is not the same thing as Two Hat.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What AI moderation can—and cannot—solve
Policy determines the result
There is no universal definition of harmful content. Results depend on a platform’s rules, selected severity thresholds, language and cultural context, and whether the system is analyzing text, images, video, audio or several modalities together.
Free tools Windows power users keep installed
One-click scans. No signup required.
False positives and false negatives remain unavoidable
- False positive: permitted content is blocked or sent for review.
- False negative: abusive or dangerous content passes through.
- Context failure: irony, quotation, counterspeech, reclaimed language or community slang is misunderstood.
- Adversarial evasion: users alter spelling, Unicode, spacing, emojis or coded language to avoid detection.
- Language imbalance: performance can vary among supported languages.
- Multimodal mismatch: an image and its accompanying text may change each other’s meaning.
Human review is still necessary
People are needed for ambiguous cases, appeals, policy exceptions, high-impact decisions, emerging abuse patterns and legally sensitive or safety-critical material. Automated systems are most useful as a scalable detection and triage layer inside a broader governance process.
Privacy claims require product-specific reading
Microsoft’s acquisition announcement included assurances about customer data and confidentiality, but that statement is not a complete technical privacy audit. For Azure AI Content Safety, Microsoft says customer prompts and completions are not stored or used to train filtering models without consent and remain in the selected Azure region during processing. Those Azure statements should not automatically be generalized to every Community Sift deployment or historical Two Hat agreement.
What remains unknown
- Microsoft has not disclosed the transaction value.
- Public sources do not provide a detailed post-acquisition integration architecture.
- There is no independent, industry-wide accuracy benchmark establishing that the system outperforms all alternatives.
- Public information does not account for every original Two Hat customer contract after the acquisition.
- Nothing in the available evidence supports claiming that the acquisition eliminated online abuse.
Why the 2021 deal still matters
Microsoft’s purchase was a strategic investment in scalable online-community safety, particularly for gaming. Its significance today is less the undisclosed corporate price than the way moderation capability has been incorporated into Microsoft’s continuing Community Sift and Xbox safety work. The acquisition also illustrates the practical limits of “AI moderation”: models can find and prioritize likely violations, but platform policies, human review and enforcement decisions determine what users ultimately experience.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
