Adobe for Slack transforms chat discussions into images, videos, and PDFs using Firefly.

      Adobe for Slack allows Slackbot to transform discussions and shared files into images, videos, and documents utilizing Firefly, Photoshop, Premiere, Acrobat, and over 70 additional tools. According to Article 50 of the EU AI Act, synthetic content must now be labeled in a machine-readable format, and Adobe’s C2PA standard is mentioned in the Commission’s guidance as a means to achieve this.

      Adobe has integrated its creative tools into Slack, with Slackbot sending prompts to Firefly, Photoshop, Premiere, or Acrobat and returning the resulting files, as reported by Digital Trends. It analyzes the channel to interpret user intentions, extracting information from conversations, canvases, and shared files, and can search the Creative Cloud library based on subject, style, or mood. More than 70 Adobe tools are linked in this feature.

      This integration is not included in standard subscriptions; users require either a Business+ or Enterprise+ Slack plan along with a separate Adobe subscription. Teams can collaborate and refine assets within channels.

      This represents the third of such integrations. Adobe made a similar integration for ChatGPT last month, already has a Claude version, and has announced plans for Gemini.

      The European issue lies in the output generated. Since August 2, 2026, Article 50 has mandated that synthetic images, videos, and audio must be labeled in a machine-readable format. Adobe is in a strong position to address this, having established the Content Authenticity Initiative and co-authored C2PA, which is the provenance standard for Content Credentials.

      Brussels has effectively endorsed C2PA, with the Commission's transparency code identifying it as a prime example meeting all four required criteria, according to some analyses. These criteria include robustness, which necessitates that marking must endure format changes and minor edits while being tamper-evident, to the extent technically possible. Interoperability is also required.

      This presents an intriguing scenario for a chat client. Manifests are preserved through conversions done by C2PA-compliant tools, but a simple re-save may erase them. It's currently unclear whether these elements remain intact through this specific process: generating in Slack, downloading, and integrating into a presentation may or may not keep the credentials attached.

      The code anticipates this issue, recommending the pairing of C2PA with an invisible watermark, a method that OpenAI has adopted with Google’s SynthID. Others are addressing the issue from a different angle; YouTube now labels AI-generated videos regardless of whether the creators disclose this information, implying rather than marking at the source.

      This presents a notable gap. Adobe authored the standard, and TNW reported that Anthropic is marking everything as the same regulations come into effect, yet there has been no testing on how a chat client may alter these elements.