TL;DR: SB 942 became operative August 2, 2026. Midjourney ships no C2PA content credentials and no pixel watermark, making it the highest-profile non-compliant AI image generator under California law. The California AG and, for the first time in US AI regulation, city attorneys and county counsel have authority to enforce the $5,000/day penalty. Teams relying on Midjourney for commercial content output face immediate exposure.
California's AI watermarking law had a clear operative date: August 2, 2026. As of that date, covered AI image and video generators must embed detectable disclosure in their outputs. Midjourney, one of the most widely used AI image platforms in the US, crossed that date with no C2PA content credentials and no visible pixel watermark.
That compliance posture makes Midjourney the clearest high-profile case of a major AI image generator that appears to be in violation of SB 942 as of its first day of enforcement.
What SB 942 requires from covered providers
The California AI Transparency Act, Senate Bill 942 as amended by Assembly Bill 853, covers AI providers whose generative AI systems attract more than one million monthly visitors or users and are publicly accessible within California. That threshold is intentionally broad -- Midjourney easily clears it.
Covered providers must implement detectable disclosure mechanisms for synthetic content they generate. The law gives providers two technical compliance paths:
C2PA content credentials. The Coalition for Content Provenance and Authenticity standard embeds cryptographically signed metadata into image files identifying them as AI-generated, recording which model produced them, and flagging any post-processing. C2PA credentials survive file transfers but can be stripped by certain image editing operations.
Visible watermarks. A visible label on the output itself disclosing AI generation. The law requires that watermarks survive common resizing and compression -- a standard that rules out fragile overlays easily removed by re-saving.
Providers do not need to implement both; either path satisfies the statute. But Midjourney currently implements neither.
Beyond provider obligations, the law also applies disclosure requirements to certain deployers who publish AI-generated content to California consumers, particularly in political advertising and news contexts. This is a separate compliance layer from the provider obligation, and it can reach organizations that are not AI developers themselves.
Where Midjourney stands
Midjourney's current output has no embedded C2PA content credentials. Images generated through its platform arrive as standard JPEG or PNG files with no provenance metadata attached. The company has not announced a compliance roadmap or an implementation timeline for SB 942.
The law's threshold is met: Midjourney has substantially more than one million monthly users, the majority of whom are in California or access the platform from California. The "publicly accessible within California" condition is also met.
This puts Midjourney in a straightforward non-compliant posture under the law's plain text. The California AG does not need to allege willful violation -- the law's civil penalty structure applies to providers who fail to implement the required disclosure mechanisms, regardless of intent.
What distinguishes the SB 942 enforcement structure from many other AI laws is who can bring claims. The California AG has primary enforcement authority, but SB 942 also grants city attorneys and county counsel independent standing to file claims and to recover attorney fees if they prevail. This fee-shifting provision makes enforcement economically viable for smaller jurisdictions that would not otherwise have resources to pursue AI compliance cases.
Midjourney has no operations in California -- it is a San Francisco-based company that nonetheless is clearly subject to the law as a covered provider serving California residents.
What the enforcement calendar looks like
SB 942 does not require a complaint or investigation trigger before penalties begin to accrue. The $5,000-per-day structure means that from August 2, 2026 forward, each day a covered provider remains non-compliant is a day of potential penalty exposure.
The California AG's office has not announced an initial enforcement action. In new regulatory regimes, agencies often allow a short observation period before filing the first case, both to gather evidence of non-compliance and to establish clean enforcement facts. That pattern does not mean penalties are paused -- it means the AG is choosing when to file.
City attorney offices in Los Angeles, San Francisco, and San Diego have each made public statements signaling interest in AI enforcement. The fee-shifting structure makes a Midjourney action financially attractive for any city attorney whose legal staff can build the case.
What compliant watermarking actually looks like
Understanding the technical bar matters for teams evaluating alternative tools.
Adobe Firefly is the most clearly compliant major AI image generator. Every image produced through Firefly carries C2PA content credentials embedded in the file metadata. Users can verify credentials through the Content Authenticity Initiative's verification tool at verify.contentauthenticity.org.
DALL-E 3, accessed through OpenAI's API or products, implements C2PA Selective Hash Credentials (SHC). The credentials are embedded at generation time. OpenAI has published documentation describing its C2PA implementation, which makes compliance auditing straightforward.
Google's AI image generation tools have more variable C2PA adoption depending on which surface is being accessed. The API and some consumer products embed credentials; others do not. Teams using Google AI image tools for production content should verify C2PA status for their specific integration.
The practical takeaway: for commercial content workflows where SB 942 compliance matters, Adobe Firefly and the OpenAI API (with C2PA enabled) are cleaner choices today than Midjourney.
5-point action plan for teams using Midjourney
If your organization uses Midjourney to produce content that reaches California consumers, these are the steps to take now.
1. Audit your Midjourney output volume and use cases. Identify where AI-generated images from Midjourney appear in your published content, advertising, social media, or product UI. Political advertising and news contexts carry the highest immediate risk under the law's deployer provisions. Marketing content carries lower immediate risk but should be in scope of your audit.
2. Apply your own disclosure labels where output is published. While SB 942's primary obligation falls on Midjourney as the provider, the deployer provisions of the law and federal disclosure standards from the FTC both expect organizations that know they are publishing AI-generated content to disclose it. A visible "AI-generated" label on published images reduces your exposure regardless of Midjourney's compliance status.
3. Evaluate compliant generator alternatives for critical use cases. For content categories where non-disclosure creates the most risk -- political advertising, news illustrations, product images subject to FTC endorsement guidance -- switch to a generator with verified C2PA credentials for those outputs. Adobe Firefly and the OpenAI API are the two most straightforward options today.
4. Document your disclosure practices. If you do publish Midjourney output, document that you added your own disclosure labels and how. Regulators in enforcement proceedings look at what the organization did to comply, not only whether the underlying tool was compliant.
5. Watch for AG guidance on the deployer provisions. SB 942's deployer obligations are less clearly defined than the provider obligations. The California AG is expected to issue guidance clarifying which deployer categories face immediate enforcement attention. Subscribe to AG enforcement communications so you do not miss the first guidance documents.
What to watch after August 2
The first enforcement action under SB 942 will set the precedent for how aggressively the AG pursues the penalty structure. A Midjourney case, given the clarity of non-compliance and the platform's name recognition, would be a high-visibility first action. But the AG may choose to start with a smaller provider where the legal facts are even cleaner.
Watch for city attorney filings in Los Angeles and San Francisco. Both offices have staffed up AI enforcement capabilities and the fee-shifting provision makes these economically rational cases for city legal teams.
Midjourney's response to the law -- whether they announce a compliance roadmap, argue that the law does not apply to them, or remain silent -- will shape how the enforcement community treats similarly situated providers. An extended silence from Midjourney will be read as an absence of good-faith compliance effort, which in most regulatory contexts is treated as an aggravating factor.
For the broader AI watermarking landscape, SB 942 is the most aggressive US state law but it is not alone. The EU AI Act Article 50 transparency obligations, which became enforceable on August 2, 2026 as well, require disclosure of AI-generated content in overlapping categories. Teams building compliance programs for one law will find significant overlap with the other.
