Collaborative coauthoring with artificial intelligence

Patent metadata

Field	Value
Publication number	US-12118296-B2
Application number	US-202318309338-A
Country	US
Kind code	B2
Filing date	Apr 28, 2023
Priority date	Mar 9, 2023
Publication date	Oct 15, 2024
Grant date	Oct 15, 2024

Abstract

A data processing system implements receiving a first textual prompt from a first user via a first user interface of a first client device, the first textual prompt includes a textual description of content to be generated by a generative model; providing the first textual prompt as a first input to the generative model to obtain as an output first collaborative content based on the first textual prompt; receiving an edited version of the first collaborative content from the first client device that includes edits to the first collaborative content provided by the first user; providing the edited version of the first collaborative content as a second input to the generative model to obtain as an output second collaborative content based on the edited version of the first collaborative content; and causing the first client device to present the second collaborative content on the first user interface.

First claim

What is claimed is:

1. A data processing system comprising:
a processor; and
a machine-readable medium storing executable instructions that, when executed, cause the processor to perform operations comprising:
receiving, during a coauthoring session between a first user and a first generative model, a first textual prompt from the first user via a user interface of a first client device, the first textual prompt comprising a first textual description;
providing the first textual description as a first input to the first generative model to output a first collaborative content including textual and non-textual collaborative content;
applying on the non-textual collaborative content of the first collaborative content at least one of an image analysis to identify a subject matter depicted therein, or an optical character recognition to extract textual components depicted therein;
monitoring the textual collaborative content and at least one of the subject matter or the textual components for objectionable content;
causing the first client device to present the first collaborative content for editing during the coauthoring session, after determining that the textual collaborative content and the at least one of the subject matter or the textual components do not contain objectionable content;
receiving during the coauthoring session an edited version of the first collaborative content from the first client device, the edited version of the first collaborative content including edits to the first collaborative content provided by the first user via the user interface;
providing the edited version of the first collaborative content as a second input to the first generative model to output a second collaborative content;
causing the first client device to present the second collaborative content on the user interface during the coauthoring session;
maintaining a history of textual prompts, a plurality of collaborative content, and edits to the plurality of collaborative content generated by the first generative model during the coauthoring session, wherein the history includes controls that, when activated, present a view of a state of content generated by the first generative model for each state, the textual prompts include the first textual prompt, and the plurality of collaborative content include the first collaborative content; and
causing the first client device to present the textual prompts and the plurality of collaborative content in a chronological list in conjunction with a current version of the collaborative content in one screen on the user interface during the coauthoring session.

2. The data processing system of claim 1 , wherein the non-textual collaborative content of the first collaborative content comprises a diagram, a drawing, an image, a video, or a combination thereof, wherein each user entry is displayed with a user profile icon of a user who authored the user entry in the screen, and
wherein the current version of the collaborative content is displayed with a coauthoring icon on the screen.

3. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
receiving a second textual prompt from the first user with the edited version of the first collaborative content;
providing the second textual prompt and the edited version of the first collaborative content as the second input to the first generative model to obtain the second collaborative content based on the edited version of the first collaborative content.

4. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
monitoring the first textual prompt, the second textual prompt, and the edited version of the first collaborative content for objectionable content; and
rejecting the first textual prompt, the second textual prompt, or the edited version of the first collaborative content for objectionable content responsive to the first textual prompt, the second textual prompt, or the edited version of the first collaborative content including objectionable content.

5. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
monitoring the second collaborative content for objectionable content;
and
discarding the second collaborative content responsive to the second collaborative content including objectionable content.

6. The data processing system of claim 1 , wherein the first generative model is a Generative Pre-trained Transformer 3 model or a Generative Pre-trained Transformer 4 model, and wherein the screen presents the chronological list before the current version of the collaborative content.

7. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
causing the first client device to present one or more action tiles in another screen on the user interface, wherein the one or more action tiles are to be activated to generate a summary of the collaborative content currently on the screen, to generate a content type other than a current content type of the collaborative content currently on the screen, or a combination thereof.

8. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
receiving a third textual prompt from a second client device of a second user;
providing the third textual prompt as a third input to the first generative model to obtain third collaborative content based on the second collaborative content and the third textual prompt;
causing the first client device and the second client device to present the third collaborative content on the user interface.

9. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
receiving a third textual prompt from a second client device of a second user;
receiving an edited version of the second collaborative content from the first client device, the edited version of the second collaborative content includes edits to the second collaborative content provided by the second user;
providing the third textual prompt and the edited version of the second collaborative content as a third input to the first generative model to obtain third collaborative content based on the edited version of the second collaborative content and the third textual prompt;
causing the first client device and the second client device to present the third collaborative content on the user interface.

10. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
receiving an indication from the first client device to activate an entry in the history; and
causing the first client device to present a version of the collaborative content when a prompt associated with the entry was submitted to the generative model.

11. The data processing system of claim 1 , wherein the machine-readable medium includes instructions configured to cause the processor to perform operations of:
receiving an indication from the first client device to split the history into multiple forks;
continuing to revise the textual content generated by the first generative model independently on each fork of the history.

12. The data processing system of claim 1 , wherein t

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Wilde Benjamin Gustav
• Richardson Ransom Lloyd
• Hong Jenna
• Lorrain-Hale Theo
• Long Christina Rose
• Tuffaha Omar Basim
• Wang Yang
• Scudieri Paul Anthony
• Duswald Susanne
• Lawrence Ryan Michael
• Scarpati Michael Adam
• Cochrane Courtney Sarah
• Romanov Alexey
• Rotella Katie Nicole
• Ghadiyali Moiz Shabbir
• Koui-Miclette Emili

Classifications

• G06F40/56 Primary

  Natural language generation · CPC title

• G06F16/176

  Support for shared access to files; File sharing support · CPC title

• H04L65/4015

  where at least one of the additional parallel sessions is real time or time sensitive, e.g. white board sharing, collaboration or spawning of a subconference · CPC title

• G06F40/166 Primary

  Editing, e.g. inserting or deleting · CPC title