System and method for persuadable collaborative conversational AI

A method for improved decision-making in a collaborative artificial intelligence having an experience chain data structure based on past experience, the steps comprising:

• receiving a conversational input in natural language form, said input to which the collaborative artificial intelligence responds by deciding on a response in collaboration with at least one other collaborative artificial intelligence;
• connecting the collaborative artificial intelligence to the at least one other collaborative artificial intelligence via a facilitator;
• recording the conversational input as a new experience chain link;
• adding at least one of associated contexts, memories, correlations, and causations to the new experience chain link;
• determining at least one desired outcome for each collaborative artificial intelligence connected to the facilitator;
• predicting a value of the at least one desired outcome for each collaborative artificial intelligence connected to the facilitator;
• selecting a response based on the predicted value of the at least one desired outcome across all collaborative artificial intelligences connected to the facilitator;
• transmitting the response to a receiver;
• observing a result of the response;
• and
• creating a new extended experience chain data structure including context, prediction, correlation, and causation links to encode reinforcement based on the observed result.

A system for persuadable artificial intelligence participating in an application conversation, the system comprising:

• a forum hosting a collaborative conversation;
• at least two subminds bi-directionally coupled to the forum, said subminds further including
• a basic intelligence to create a proposed response for the application conversation,
• a natural language understander to prepare and process proposed responses from the at least two subminds to participate in a selection of a response in the collaborative conversation, and
• a conduit bi-directionally coupled to the forum, bi-directionally coupled to the basic intelligence, and bi-directionally coupled to the natural language understander;
• wherein at least one of the bi-directionally coupled subminds further includes a Decidron having an evaluator coupled to the conduit and a database coupled to the Decidron and to the natural language understander, where the Decidron creates and stores experience chain data structures to the database and retrieves said experience chain data structures for use in decision-making when collaborating with other subminds bi-directionally coupled to the forum;
• and
• at least one facilitator, bi-directionally coupled to the forum and to the application conversation, to proctor protocols in the collaborative conversation, and to reconvey a prompt and response selected by the coupled subminds from and to the external application conversation.

A method for improved decision-making in an artificial intelligence having an experience chain data structure based on past experience, the steps comprising:

• receiving an input in natural language form;
• recording the input as at least one experience chain element;
• linking the at least one experience chain element to other experience chain elements from the experience chain data structure representing at least one of associated contexts, memories, correlations, and causations;
• determining at least one desired outcome;
• predicting a value of the at least one desired outcome;
• constructing candidate responses based on the associated contexts, memories, correlations, and causations from the experience chain data structure;
• selecting a response from the candidate responses based on the predicted value of the at least one desired outcome;
• transmitting the selected response to a receiver:
• observing a result of the response;
• and
• creating an extended experience chain data structure including context, prediction, correlation, and causation links to encode reinforcement based on the observed result.

A method for improved decision-making in a collaborative artificial intelligence having an experience chain data structure based on past experience, the steps comprising:

• receiving a conversational input in natural language form;
• connecting the collaborative artificial intelligence to at least one other collaborative artificial intelligence via a facilitator;
• recording the conversational input as at least one experience chain element:
• linking the at least one experience chain element to other experience chain elements from the experience chain data structure representing at least one of associated contexts, memories, correlations, and causations;
• determining at least one desired outcome;
• predicting a value of the at least one desired outcome;
• constructing at least one candidate response for each of the connected collaborative artificial intelligences based on the associated contexts, memories, correlations, and causations from the experience chain data structure;
• selecting a response to the received conversational input from the at least one candidate response based on the predicted value of the at least one desired outcome;
• transmitting the selected response to a receiver;
• observing a result of the response;
• and
• creating an extended experience chain data structure including context, prediction, correlation, and causation links to encode reinforcement based on the observed result.

A system for persuadable artificial intelligence participating in an application conversation, the system comprising:

• a forum hosting a collaborative conversation;
• and
• at least two subminds bi-directionally coupled to the forum;
• wherein at least one of the bi-directionally coupled subminds further includes a machine learning unit having an evaluator, where the machine learning unit creates and stores experience data structures to a database and retrieves said experience data structures for use in decision-making when collaborating with other subminds bi-directionally coupled to the forum.