Under Community Review

Currently building generative AI capabilities in upcoming Product Releases to support both authoring and delivery use cases. 

ChatGPT Connector for Tridion Sites & Tridion Docs for Content Generation

A struggle many business teams face is generating and delivering content quickly.  The most time consuming challenge is "starting."  It’d be great if business teams could tap into OpenAI / ChatGPT while authoring content in Tridion Sites & Docs, leveraging it directly in Tridion to aid in content generation.

Jim Rabon
Parents

  • So would this directly populate into Tridion? One would think there would be content policies - external copy / copywriter approved per anything generated by ChatGPT and published - even to lower environment's. Maybe there is not governance with your company?

  • in reply to Andrew Ross

    Good point. For AI generated content, we’re already seeing concerns from organizations especially around copy-writable content where organizations do not allow any AI-generated content for copy-writable material. I’m not a lawyer, but I understand the thinking is AI-created works (supposedly) don’t have the same copyright protection as things created by humans.

    For the use case for the product, a good fit would be generative AI relying on the organization’s existing data, under the organization’s policies around AI for sure. Also, the content would (or should) be treated as recommendations with final authoring and approval by humans.

    For example, the semantic tagging feature for Tridion Sites and Tridion Docs can propose tags based on the content in the Components or Topics. This could be based on a generic model or tuned to the organization’s data. Ultimately, the authors will review, change, and check in the final tags or keywords.

    One more use case that’s already proven in the industry is machine translation. Some content can be automatically translated with caveats on the quality, but we also have the scenario of machine translation with post edits by humans. AI technology helps, but ultimately humans make the final content or translation in this case. I think the use case for generative AI will depend on the organization and the type content as well.

Comment
  • in reply to Andrew Ross

    Good point. For AI generated content, we’re already seeing concerns from organizations especially around copy-writable content where organizations do not allow any AI-generated content for copy-writable material. I’m not a lawyer, but I understand the thinking is AI-created works (supposedly) don’t have the same copyright protection as things created by humans.

    For the use case for the product, a good fit would be generative AI relying on the organization’s existing data, under the organization’s policies around AI for sure. Also, the content would (or should) be treated as recommendations with final authoring and approval by humans.

    For example, the semantic tagging feature for Tridion Sites and Tridion Docs can propose tags based on the content in the Components or Topics. This could be based on a generic model or tuned to the organization’s data. Ultimately, the authors will review, change, and check in the final tags or keywords.

    One more use case that’s already proven in the industry is machine translation. Some content can be automatically translated with caveats on the quality, but we also have the scenario of machine translation with post edits by humans. AI technology helps, but ultimately humans make the final content or translation in this case. I think the use case for generative AI will depend on the organization and the type content as well.

Children
No Data