Project Summary:
The Hypermanifest is a novel framework designed to facilitate mediated human/AI interaction across different “time zones.” By dividing interactions into Immediate Response, Short Term, and Long Term thinking, the Hypermanifest introduces nuanced handling of inquiries and requests, allowing for a more sophisticated collaboration between humans and AI.
This project will develop an interface prototype that will eventually evolve into a fully fleshed product. The development process will include essays, reviews, feedback sessions, and conference presentations/papers to document and refine the concept.
My current research on the Hypermanifest can be found here:
Introduction
Time Mediation
Reimagining Wisdom
Initially, the interface will be text-based with future plans to integrate Brain-Computer Interfaces (BCI), pushing the boundaries of AI-human interaction. While there is a push towards fully automated research that highlights the growing capabilities of AI, the Hypermanifest emphasises the importance of continued human participation. The structure of collaborative, time sensitive interactions between humans and AI ensures that human insights and ethical consideration remains central - complementing the potential of automation and not replacing it.
Here’s an example application of the Hypermanifest:
Climate Change Mitigation Research
Initial Login
The user logs in (potentially through biometric or BCI verification), selecting their focus areas (e.g. Urban Emissions, Arctic Sea Ice)
Interface adapts via personalised dashboard, showing AI analysis and data streams
Input Stream
Researcher begins a thought stream (either through a BCI or advanced natural language interface) e.g. “need to compare with historical data/project future scenarios on rapid ice melt”
System receives input and visually represents the thought across time zones
Hypermanifest Activation: Examples across all 3 Time Zones
Immediate Response:
Dynamic Visualisations showing real time changes in ice coverage
The flagging of unusual weather patterns
Highlighting anomalies in Arctic sea ice extent data
Short Term Zone
The system prompts the user to input their interpretations, suggesting relevant factors as they think or type - adapting to the researcher’s train of thought
The interface shows a collaborative/latent space where other insights from experts are shared.
System highlights potential conflicts/synergies between opinions
Long Term Zone
Focuses on multi-model projections and generations of future scenarios e.g. Arctic ice extent, considering various climate models
Highlighting potential tipping points and effects on global climate with probability assessments
Suggesting long term research directions and policy interventions based on analysis.
The aim is to be able to switch between these seamlessly, creating a latent space by which humans and AI are on the same page.
What makes the Hypermanifest stand out?
Contains an adaptive interface based on user interaction patterns
Ethical AI layer ensuring transparency in both model assumptions and data sources
Dynamic knowledge presentation - visualisation the interplay between human intuition/AI analysis
Intuition amplification system - learning from and enhancing human insights
What are this project's goals? How will you achieve them?
The Hypermanifest aims to:
Enhance collaborative efforts between humans and AI by aligning insights and workflows more closely, similar to but more intensive than the Human-in-the-Loop (HITL) model.
Introduce a time-based system for AI and human collaboration, utilising the Global Workspace System to analyse sentiment and time-sensitive language, thereby integrating human intuition (as vectors) and AI’s data-driven insights (as data manifolds).
Value human contribution while complementing the potential of automation.
I will achieve these goals by:
Developing a functional prototype that demonstrates the concept.
Conducting extensive user testing and feedback sessions to refine the interface.
Presenting findings at conferences and through published papers, engaging with the broader research community to validate and improve the system.
How will this funding be used?
I’m currently seeking $50,000 in funding to cover the following:
Prototyping: Funding will cover compute resources and programming costs to develop the initial interface ($20,000)
Researcher Salary: To support dedicated research and development time ($15,000)
Conference Attendance: Registration fees, accommodation, and travel expenses for presenting the project at key conferences ($10,000)
Misc. Expenses: Publication fees, Community Engagement activities, any surprise project-related expenses ($5,000)
Who is on your team? What's your track record on similar projects?
Dr. Suzie Gray - Independent Researcher:
AI Education Content Creator on Substack, Instagram and TikTok
PhD in Science Fiction Theatre, with a strong interdisciplinary background
Founder of GLIMPSE, a futurist group that facilitated collaboration between artists and scientists
Experience with Lunar Mission One collaboration projects
Developer of Poetal Lite, an AR poetry app under indie games company Critical Lit Games
Created AR art for Virtual Futures in London
What are the most likely causes and outcomes if this project fails?
Causes of Failure:
Technical Challenges: Integrating Human intuition with AI seamlessly can prove difficult, especially with the addition of BCIs.
Lack of Engagement: If the project fails to capture the attention of the industry/community, to which we could build iterative prototypes with potential A/B testing
Outcomes of Failure:
Learning and Development: Even if the project succumbs to the above failures, the R&D process will yield valuable insights that can inform future projects.
Partial Success: Specific parts of the project, like the time-based analysis system, could be adapted for other applications.
What other funding are you or your project getting?
At this stage, the project is primarily self-funded, but I’m currently seeking additional funding from relevant AI and tech grants.
v1.1 - 15/08/24