Methods and systems for enabling collective superintelligence
The present disclosure describes systems and methods for enabling real-time conversational dialog among a large population of networked human users while facilitating convergence on groupwise decisions, insight extraction, and solutions, and amplifying collective intelligence. A plurality of computing devices associated with members of a population of participants is provided, wherein the plurality of computing devices is networked via a computer network and in communication with a central server. In some cases, the central server divides the population into a plurality of subgroups and enables an LLM-powered conversational surrogate agent associated with each subgroup. The conversational surrogate agent is configured to repeatedly observe a conversation among members of its assigned subgroup and store content of the conversation to extract insights from the conversation at appropriate intervals. The conversational surrogate agent assesses a numerical measure of conviction, confidence, or scope associated with each unique insight within the conversation, passes the insights and the numerical measure to other subgroups and modulates a strength of the insights based on the numerical measures.
1. A system for real-time communication, conversation analysis and insight extraction, comprising:
a plurality of networked computing devices associated with members of a population of participants, and networked via a computer network; and
a central server in communication with the plurality of networked computing devices, the central server dividing the population into a plurality of subgroups and enabling a conversational surrogate agent (CSai) associated with each subgroup, each conversational surrogate agent comprising a large language model (LLM)-powered module that is assigned to a unique subgroup of the population of participants and tasked with repeatedly performing the following steps in real-time, comprising:
observing via the computer network a conversation among members of its assigned subgroup,
distilling and storing content of the conversation,
extracting insights from the conversation at intervals,
assessing a numerical measure of conviction, confidence or scope associated with each unique insight within the conversation among the members of the assigned subgroup,
storing the insights and associated numerical measures and associated members in a memory by passing data to one or more other computational processes accessible to the central server,
aggregating the associated numerical measures across unique insights expressed within the assigned subgroup,
passing insights to other conversational surrogate agents assigned to other subgroups, said insights being passed as language, each insight being passed along with said associated numerical measures having been aggregated,
receiving insights from the other conversational surrogate agents assigned to other subgroups, each insight including textual language representing the insight and aggregated numerical measures associated with the other subgroups, and
conversationally expressing received insights to the assigned subgroup, said expressing including modulating a strength of the textual language based at least in part on the received aggregated numerical measures.
2. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent performing the following additional step in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroup or subgroups within the connection set will receive the insight by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time,
wherein the determining which insight to transmit comprises determining which insight has the highest support within the assigned subgroup.
3. The system of claim 2 wherein the determining which other subgroup or subgroups within the connection set will receive the insight further includes considering which of the other subgroups within the connection set have not yet considered the insight.
4. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent periodically performing the following additional step on a prescribed period in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroup or subgroups within the connection set will receive the insight by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time,
wherein the determining which insight to transmit comprises determining which insight has the highest support within the assigned subgroup, and
wherein the prescribed period varies as a function of how high support is for the insight within the assigned subgroup.
5. The system of claim 4 wherein the determining which other subgroup or subgroups within the connection set will receive the insight further includes considering which of the other subgroups within the connection set have not yet considered the insight.
6. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent periodically performing the following additional step on a prescribed period in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroup or subgroups within the connection set will receive the insight by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time,
wherein the determining which insight to transmit comprises determining which insight has the highest support within the assigned subgroup, and
wherein the prescribed period shortens as a level of support increases for the insight within the assigned subgroup.
7. The system of claim 6 wherein the determining which other subgroup or subgroups within the connection set will receive the insight further includes considering which of the other subgroups within the connection set have not yet considered the insight.
8. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent periodically performing the following additional step on a prescribed period in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroups within the connection set will receive the by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time, and
determining at said moment in time whether to set a ready to send state as a function of whether the insight determined to transmit from said assigned subgroup is a minority opinion amongst insights for each of the plurality of subgroups.
9. The system of claim 8 wherein said assigned subgroup will not send said determined insight to said determined subgroup or subgroups unless said ready to send state is set.
10. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent periodically performing the following additional step on a prescribed period in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroups within the connection set will receive the insight by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time, and
determining at said moment in time whether to set a ready to send state as a function of an amount of time since the ready to send state was set.
11. The system of claim 10 wherein said assigned subgroup will not send said determined insight to said determined subgroup or subgroups unless said ready to send state is set.
12. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent periodically performing the following additional step on a prescribed period in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroups within the connection set will receive the by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time, and
determining at said moment in time whether to set a ready to send state as a function of an amount of conversational content that has been generated among members of the assigned subgroup since the ready to send state was set.
13. The system of claim 12 wherein said assigned subgroup will not send said determined insight to said determined subgroup or subgroups unless said ready to send state is set.
14. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent periodically performing the following additional step on a prescribed period in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroups within the connection set will receive the insight by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time, and
determining at said moment in time whether to set a ready to send state as a function of whether the insight determined to transmit from said assigned subgroup has a confidence, conviction, and/or scope exceeding a prescribed threshold value.
15. The system of claim 14 wherein said assigned subgroup will not send said determined insight to said determined subgroup or subgroups unless said ready to send state is set.
16. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent performing the following additional step in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and which other subgroups within the connection set will receive the insight by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time,
wherein the determining which insight to transmit comprises determining a set of insight that have the highest support within the assigned subgroup, and randomly selecting one of the insights from the set of insights.
17. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent performing the following additional steps in real time, comprising:
determining at a moment in time which insight to transmit from said assigned subgroup and to which other subgroups within the connection set will receive the insight by considering which of the other subgroups of the connection set are currently in a ready to receive state at said moment in time, and
selecting via a mechanism one or more said other subgroups of said connection set that are in the ready to receive state to receive the insight.
18. The system for real-time communication, conversation analysis and insight extraction of claim 17 further comprising:
said central server in communication, said conversational surrogate agent performing the following additional step in real time, comprising:
said selecting via said mechanism, wherein said mechanism comprises prioritizing for receiving the insight said one or more said other subgroups that have not yet discussed the insight.
19. The system for real-time communication, conversation analysis and insight extraction of claim 17 further comprising:
said central server in communication, said conversational surrogate agent performing the following additional step in real time, comprising:
said selecting via said mechanism, wherein said mechanism comprises prioritizing for receiving the insight said one or more said other subgroups that have a lowest level of support for the insight.
20. The system for real-time communication, conversation analysis and insight extraction of claim 17 further comprising:
said central server in communication, said conversational surrogate agent performing the following additional step in real time, comprising:
said selecting via said mechanism, wherein said mechanism comprises prioritizing for receiving the insight said one or more said other subgroups that have not yet discussed the insight and in the event that there are none of said other subgroups that have not yet discussed the insights, prioritizing for receiving the insight said one or more said other subgroups that have a lowest level of support for the insight.
21. The system for real-time communication, conversation analysis and insight extraction of claim 17 further comprising:
said central server in communication, said conversational surrogate agent performing the following additional step in real time, comprising:
said selecting via said mechanism, wherein said mechanism comprises prioritizing for receiving the insight said one or more said other subgroups as a function of random selection.
22. The system for real-time communication, conversation analysis and insight extraction of claim 21 further comprising:
said selecting wherein said selection via said mechanism, wherein said mechanism comprises prioritizing for receiving the insight said one or more said other subgroups as a function of random selection amongst said one or more said other subgroups that have a lowest level of support for the insight.
23. The system for real-time communication, conversation analysis and insight extraction of claim 21 further comprising:
said selecting wherein said selection via said mechanism, wherein said mechanism comprises prioritizing for receiving the insight said one or more said other subgroups as a function of random selection amongst said one or more said other subgroups that have a lowest level of support for the insight.
24. The system for real-time communication, conversation analysis and insight extraction of claim 21 further comprising:
said selecting wherein said selection via said mechanism, wherein said mechanism comprises prioritizing for receiving the insight said one or more said other subgroups as a function of random selection amongst said one or more said other subgroups that have a particular range of support levels for the insight.
25. The system for real-time communication, conversation analysis and insight extraction of claim 1 further comprising:
said central server in communication, the central server identifying for each of the plurality of subgroups a connection set comprising a set of other subgroups that the assigned subgroup can exchange insights with via transmission of insights, said conversational surrogate agent performing the following additional step in real time, comprising:
receiving feedback on insights passed to other conversational surrogate agents assigned to the other subgroups.
26. The system of claim 1 , further comprising said assessing said numerical measure of conviction, confidence or scope associated with each unique insight within said conversation among said members of the assigned subgroup, wherein the numerical measure of conviction, confidence or scope is calculated based on a sentiment data structure of each individual in the population of participants.
27. The system of claim 1 , further comprising said assessing said numerical measure of conviction, confidence or scope associated with each unique insight within said conversation among said members of the assigned subgroup, wherein the numerical measure of conviction is calculated as a fraction of positive sentiment directed towards each insight that has been conversationally deliberated within the assigned subgroup during a period of time.
28. The system of claim 27 , wherein positive sentiment of an individual in an insight represents a degree to which that individual believes that insight is a good answer to a question or issue being deliberated.
29. The system of claim 27 , wherein positive sentiment is calculated and stored in real-time for each individual in the population of participants.
30. The system of claim 27 , wherein a natural language processing system is used to transform a batch of an individual's dialog into an estimation of a degree to which the individual supports each insight.
31. The system of claim 27 , wherein positive sentiment is calculated on a zero-centered scale, wherein a center of the scale is equivalent to not having mentioned that insight.
32. The system of claim 27 , wherein a new sentiment data structure is applied to an individual's existing sentiment data structure using heuristics, comprising a combination of the following rules: (i) if the insight already exists in the existing sentiment data structure and exists in the new sentiment data structure, apply smoothing by taking a weighted average of a new sentiment value of the insight and an existing sentiment value of the insight, (ii) if the insight does not already exist in the existing sentiment data structure and exists in the new sentiment data structure, increase the existing sentiment value by some fraction of the new sentiment value, and (iii) if the insight already exists in the existing sentiment data structure and does not exist in the new sentiment data structure, apply some decay to the existing sentiment value.
33. The system of claim 1 , further comprising said assessing said numerical measure of conviction, confidence or scope associated with each unique insight within said conversation among said members of the assigned subgroup, wherein the numerical measure of scope is calculated as a fraction of users who have mentioned or referenced the insight at least once, as identified by a natural language processing tool or a sentiment data structure.
34. The system of claim 1 , further comprising said assessing said numerical measure of conviction, confidence or scope associated with each unique insight within said conversation among said members of the assigned subgroup, wherein the numerical measure of conviction, confidence or scope is calculated at an individual level, averaged for each subgroups and reported as that subgroup's conviction, confidence or scope for each insight, or averaged over the population of participants and reported as a global conviction, confidence or scope for each insight.
35. The system of claim 1 , further comprising said assessing said numerical measure of conviction, confidence or scope associated with each unique insight within said conversation among said members of the assigned subgroup, wherein said assessing said numerical measure of conviction, confidence or scope associated with each unique insight within said conversation among said members of the assigned subgroup further comprises assessing a numerical measure of exposure associated with each unique insight within said conversation among said members of the assigned subgroup, wherein the numerical measure of exposure is calculated as a fraction of users that an idea has been conversationally presented to, as identified by a natural language processing tool or a sentiment data structure.