Adobe has released a new premium AI assistant to assist in document workflows and new generative AI video tools for its Adobe Premiere Pro product. The first service, known as Acrobat AI Assistant, was previously announced in February and is being primarily marketed by the design software giant as a document summarization tool. The generative AI services for Premium Pro, meanwhile, will afford users the ability to streamline the video editing process by expediting object addition and removal, adding new frames or generating entirely new footage.
“Adobe is reimagining every step of video creation and production workflow to give creators new power and flexibility to realize their vision,” said Ashley Still, senior vice president of Adobe’s Creative Product Group. “By bringing generative AI innovations deep into core Premiere Pro workflows, we are solving real pain points that video editors experience every day, while giving them more space to focus on their craft.”
New Adobe AI product launches in Acrobat and Premiere Pro
Acrobat AI Assistant, said Adobe, will be available as part of a monthly add-on subscription of $4.99 to Adobe Acrobat. As well as summarising sections or the entirety of documents, Adobe also claims it can produce intelligent citations to allow users to more easily verify any answers the chatbot provides. Possible applications for the Acrobat AI Assistant included its use as a study guide, a summarisation tool for sales teams or a means for the wider public to spot any unusual provisos contained within long “terms and conditions” documents.
Simultaneously, Adobe also claimed that new AI innovations within its Premiere Pro platform will expedite the video editing process for consumers (no release date was specified.) Additionally, the design software firm said that editors may soon be able to leverage third-party integrations with models built by Runway or Open AI to create B-rolls for their projects. “While much of the early conversation about generative AI has focused on a competition among companies to produce the “best” AI model,” said Adobe, it added that it embraces a future wherein “thousands of specialised models emerge, each strong in their own niche.”
AI trained on AI
These two announcements come days after Bloomberg reported that the design software firm had partially trained its Firefly image generation model on images created by rival platform Midjourney. The news has proven problematic for the design software giant, with the company having explicitly marketed its Firefly platform as an ethical alternative to Midjourney in having been trained on copyright-expired public domain content and open-licensed imagery instead of data indiscriminately scraped from the wider internet.
Adobe downplayed the revelation, stating that artificially generated images constituted only 5% of the data contained in its Firefly training set. “Every image” submitted, a company spokesperson told Bloomberg, “goes through a rigorous moderation process to ensure it does not include IP, trademarks, recognisable characters or logos, or reference artists’ names.”
Read more: Adobe cancels $20bn Figma deal after regulatory pressure
