Agent-based AI
Wunder is has a deep interest in leveraging the power of AI to deleiver a more authentic social experience for our users. We beleive that leveraging agent-based AI is the most powerful way fo doing this and have aligned ourselves with a strategic partner to exclusive license their IP and evolve our own IP based on their leading agent-based AI foundations.
Agent-based AI refers to a type of artificial intelligence (AI) that uses autonomous entities, or "agents," which interact with their environment and with other agents in order to achieve specific goals or solve problems. These agents can represent software programs, robots, or virtual entities within a simulation. The agents act based on a set of rules, behaviors, or learning algorithms, and they can make decisions, adapt to changes, and sometimes collaborate or compete with other agents.
Agent-based AI can be leveraged in various ways to enhance, personalise, and optimise the user experience on Wunder. Below we have outlined ways in which we will be using agent based AI within Wunder:
Building a healthy algorithm
Agents are personalised curators: On Wunder, each user will be considered an "agent" that interacts with the platform’s content (posts, images, videos, etc.). The platform’s AI agents will analyse users' behaviours, preferences, and engagement patterns to tailor content recommendations.
Learning from interactions: The agent learns what type of content (posts, videos, ads) a user likes, comments on, shares, or ignores. Over time, the agent adapts its recommendations based on this feedback, optimising the content shown to each individual.
Collaborative filtering: Agents will work together by comparing behaviour across similar users to recommend content that others with similar tastes have engaged with.
User behaviour modelling
Simulating user interactions: In Wunder, agents will be used to model user behavior (e.g., posts, likes, comments, shares) to understand how content spreads across the platform and predict viral trends. By simulating different scenarios, the platform will predict which posts might become popular based on historical interaction patterns. We will use this information to assist creators in understanding how to improve their content and increase their reach on Wunder.
Emergent social behaviour: Agents will simulate and track emergent social behaviours like group dynamics, the formation of online communities, or the spread of misinformation, helping the platform understand social trends or detect potential issues before they escalate.
Moderation and content filtering
Automated moderation agents: Agents will be responsible for flagging inappropriate content like hate speech, spam, or explicit material by using natural language processing (NLP) or image/video recognition.
Collaborative moderation: Agents will collaborate with each other or learn from a combination of user reports, community feedback, and behavioural patterns to decide whether content should be flagged, removed, or downvoted.
Adaptation to New Trends: As new slang or emerging topics develop within user-generated content, agents will be be trained to recognise evolving patterns of inappropriate content, adapting quickly to changes in how users communicate.
Social dynamics and network analysis
Simulating social networks: By representing users and their connections as agents within a social network, you can simulate and predict how changes in the platform’s design or policies might affect social interactions, group formation, and content virality. This will enable us to build a better UX for our users and evolve the adoption of web 3.0 technologies in the future.
Identifying Influencers and Communities: Autonomous agents will analyse the behaviour of users within Wunder to identify influencers, micro-communities, or emerging trends. These insights will be used to improve user experiences and how Wunder evolves the platform to give the user a better experience, helping influencers bring even more amazing content to the people that most want it.
Echo Chambers and Filter Bubbles: Autonomous agents will help identify and mitigate the formation of echo chambers or filter bubbles where users are only exposed to content that aligns with their existing views. This might involve encouraging diverse content exposure or designing algorithms that expose users to broader perspectives. This will enable Wunder to enhance our healthy algorithm to continue to promote positive and authentic content experiences.
Ad targeting and optimisation
Targeted advertising: Advertising agents can dynamically adjust the advertisements shown to users based on their past interactions, preferences, and predicted behavior. These agents learn which ads generate the most engagement (clicks, purchases, etc.) and refine their targeting strategies. This will ensure that our users only see the advertisements most relevant to them, and ensure advertisers know the best way to reach out to our community to achieve the most success.
Auction-based Ad systems: Agents can participate in auctions to optimise bids for advertisers based on user engagement patterns. Each ad campaign could be treated as an agent that competes for the best spot based on real-time performance. This will ensure the platform is not flooded by ads that are overpowering a users feeds and a user has a fair opportunity to consume a wider array of ads based on their interests.
Sentiment analysis
Emotion-based recommendations: Agents will be used to assess the sentiment of users (via their posts or comments) and recommend content that matches their emotional state. For example, if an agent detects that a user is feeling down, it might recommend uplifting or motivating content. This feeds into our desire to be a positive and welcoming place to enjoy social media and not be a platform that supports users in spiralling into uncontrollable depressive states that other platforms do.
Community sentiment monitoring: By analysing the sentiments of large groups of users, agents will detect early signs of social unrest, viral movements, or other community-driven phenomena. This will help Wunder proactively manage the environment, responding to emerging trends like viral movements or controversial issues.
Influence and opinion dynamics
Behavioural influence models: By using agent-based modelling, Wunder can better understand how individual behaviours influence the collective behaviour of Wunder. This is particularly useful for studying viral trends, how memes spread, or how opinions shift within certain communities. This enables us to make more informed decisions about how we can continue to grow and evolve the platform.
Information diffusion: Agents will simulate how information (or misinformation) spreads through Wunder. By understanding the dynamics of information propagation, Wunder can design systems that either promote certain types of content (e.g., fact-checked news) or limit harmful behaviours like fake news dissemination.
Wunder is excited about leveraging the power of agent-based AI to do good, and to protect our community and the experience they have on a day-to-day basis.
Last updated