Meetup #77:
AI Signals 29th meetup @
Newspeak House
Tuesday, 17th March 2026
(
#proofofnetwork)
Ana Catarina De Alencar highlighted the rise of emotional AI, showing how systems that can “read" feelings are creating a new kind of relationship between humans and machines; one that feels personal, constant, and even comforting. She explained how these systems don’t just process what we say, but infer who we are through tone, behaviour, and interaction patterns, building what she described as “digital portraits” based on our emotions, insecurities, and aspirations rather than just our clicks or purchases. This shift from a behavioural data economy to a relational one raises serious questions: do people truly understand what they’re sharing, how it’s being interpreted, and what the system remembers about them over time? She highlighted growing concerns around privacy, consent, and emotional dependency, especially as younger users increasingly turn to AI companions for support and identity exploration. Most importantly, she challenged the audience to think about control: who owns these AI-generated interpretations of us, and what rights should we have to access, correct, or delete them? A closing thought to mull over: As AI becomes better at understanding and responding to our emotions, the connection may deepen, but protecting our agency will matter more than ever.
Will Chan demonstrated a fascinating idea: an AI-powered “digital twin” that can present, teach, and answer questions entirely on its own, in real time, without scripts or recordings. His system combines a 3D avatar, voice synthesis, and intelligent agents, so it can control slides, speak naturally using his own cloned voice, and respond to audience questions by pulling from a knowledge base. Visually, it’s powered by a game-like environment with a lifelike avatar that moves and lip-syncs convincingly, making the experience feel more like interacting with a person than a chatbot. What really stood out was the live Q&A, where the system could detect raised hands, pause the presentation, respond intelligently, and then continue seamlessly. Will’s bigger vision is that this is just the beginning; in the future, we could all have personal AI twins that understand our preferences, remember what we know, and even act on our behalf.
Jethro Reeve showed his personal project “Explore the Kingdom”; a powerful interactive tool that lets you explore the UK’s social and political landscape, from house prices and living costs to election results and inequality across all constituencies. Behind the scenes, it brings together 15 government data sources into one clean map, using smart data processing and visualisations like heat maps to make complex information easy to understand. His key takeaway was that while modern AI tools can now act like capable data engineers, they still need strong guardrails, because agentic systems are powerful, but not yet foolproof.