“Global Brain Network” — High-Level Project Proposal
In this document, Dr. Ben Goertzel outlines his idea for a collaborative DAIA project: to create a system (the “Global Brain Network”) into which ordinary peoples’ data and computing power can be contributed, and then systematically utilized to work on problems and questions democratically chosen by the contributing people. We call our members and community to contribute with other ideas, improvements, discussions and variations!
Brief Notes by Ben Goertzel. First version May 21, 2018. This version April 6, 2019
HIGH LEVEL GOAL:
To cause the emergent “global brain” of the planet to grow in a direction guided by the democratic and richly participatory collaboration of a wide sample of human and AI systems — rather than being dominated by the interests and wills of a few elites.
PROXIMAL GOALS:
1) To create a system (the “Global Brain Network”) into which ordinary peoples’ data and computing power can be contributed, and then systematically utilized to work on problems and questions democratically chosen by the contributing people.
2) To foster collaboration and communication among diverse developers and thinkers operating at the intersection of AI and blockchain, especially those involved with member organizations of the Decentralized AI Alliance (DAIA). Nothing builds cooperation and mutual understanding better than actually working together on something big and meaningful.
CURRENT SITUATION WITH AI ON THE PLANET:
• Right now, individual people contribute their data to big tech companies according to a poorly-accounted and poorly-defined bargain, in which the big tech companies use the data for whatever they want (in large part advertising), and people in exchange get free use of certain software products
• The data shared by people with big tech companies is commonly shared on the back end with government intelligence agencies. In China this is plain and open; in the West it is done in ways that are secretive and hard to track
• The profit big companies make from advertising services is used to power distributed AI server farms that analyze peoples’ data — essentially focusing on analyses whose results will somehow bring more ad revenue
• If AGI arises as an extension, outgrowth or emergence from current largescale AI networks, the odds seem high that this AGI will be centrally concerned with the same things as these AI networks: spying (surveillance), selling (aka ad companies like Google), killing (ah, sorry, “national defense”), and increasing wealth inequality (aka investment management for those who need it). Of course preventing terrorism, recommending products and managing investments are useful and valuable things, but should they be the main focus of our narrow AI systems today, or the AGI systems that will grow out of these tomorrow?
DESIRED SITUATION WITH AI ON THE PLANET:
• People contribute their data to a global data pool, but in a traceable and accountable way so that their data is used in ways that they approve and support and know about, in as much as possible
• When direct financial profit is obtained from a person’s data, that person should get an opportunity to share in that profit appropriately
• Profit obtained from collectively analyzing peoples’ data, will partly be used to fund the devotion of computing processor time to democratically suggested goals
• Additional processor time may be donated to the pool used to pursue these goals
• Flexible “liquid democracy” mechanisms are used to make the democratic decision of what goals to think about
• As AGI emerges from the narrow AI networks of today, it is concerned in substantial measure with the broad benefit of humans and other sentient beings, rather than with narrower commercial or national goals
TYPICAL EXAMPLE GOALS THAT THE WORLD’S BIGGEST AI SERVER FARMS ARE THINKING ABOUT RIGHT NOW:
• What video-ads should Bob Smith be shown before he watches a music video with scantily clad women in it, in order to maximize the odds that he purchases something from a particular online store?
• What news items should Jane Jones be preferentially shown in her news feed, to maximize the odds that she spends more time on the website showing the news, and clicks on more ads on this website?
• Has Ahmad been communicating with people who have expressed opinions in strong opposition to the current government of this country?
TYPICAL EXAMPLE GOALS THAT THE WORLD’S BIGGEST AI PROCESSING NETWORKS WOULD MORE DESIRABLY BE THINKING ABOUT:
• What are the genetic, biological and lifestyle patterns that distinguish people with long, healthy and happy lives from people with other sorts of lives?
• What news should Jane Jones be preferentially shown in her news feed, to increase the odds that she forms an accurate view of the world, makes decisions that are more to the benefit of herself and the world, and has a more satisfying and productive life?
• What scientific research projects should receive more funding or volunteer activity, in order to maximally militate toward the benefit of humanity and other sentient beings?
• What connections could be made between radically disconnected people in the world, so as to provide maximum increase in well-being among the connected people?
• What information is Bob Smith not aware of, that would be of most value in his life and lead to the most satisfaction and positive growth for him?
• What should the initial value system of the first really powerful AGI systems be?
• How should AGI value systems and human value systems best remain closely and productively coupled together post-Singularity, as both AGIs and humans evolve in complex and diverse ways?
USER FEATURES:
From the user point of view, the Global Brain Network would consist of an (free and open source software) app to be installed on the smartphone and/or computer, possessing the following features (among potentially others):
• Ability to “opt in” to on goingly auto-upload specific types of your data to a global democratic data pool (e.g. photos, videos, key generic words used in one’s text messages, geographical movements)
• Ability to specify certain of your data as private, not for uploading (e.g. only auto-upload photos from certain folders)
• Ability to specify a certain amount of your hard drive, RAM or processing power for potential use by the Global Brain Network
• Ability to view a list of projects that want to use specific types of your data or processing, and to approve or not (some of these projects might want to use the data for free, and some might offer payment)
• Ability to propose projects for consideration by the democratically voting community (free or for-profit)
• Ability to vote on proposed projects, and to (“liquid democracy”) delegate your votes on certain types of projects to some other person (or some AI agent)
GLOBAL BRAIN CRYPTOCURRENCIES:
For products that involve payment in exchange for use of data or processing power, payment will take place using any of multiple approved cryptographic tokens:
• Liquid utility tokens associated with DAIA member projects: As these are liquid tokens, it is possible to exchange them back and forth for other currencies such as ETH, BTC or USD
• The EquiToken, or Equitable Token: These are special, custom illiquid tokens that can only be used within the Global Brain Network software system. So payment received in EquiTokens, can only be used to get extra votes regarding what questions the Global Brain should think about. An EquiToken cannot be exchanged for any other sort of token, and cannot be transferred to any other individual.
Each user can choose to make their data and processing power available for either DAIA member project Tokens or EquiTokens at any point in time. They may also choose a certain balance, say 30% of various DAIA member Tokens and 70% EquiTokens on average over time.
The use of DAIA member project Tokens here provides a way for individual users to monetize their data and processing in various ways, in exchange for utilization of their data and processing for tasks they approve of. EquiTokens provide a way for individual users to donate their data and processing to causes they find valuable, receiving in exchange additional influence over what the Global Brain Network’s overall data and processing power are used for. The Global Brain Network’s data and processing will be apportioned in three different ways:
1. Some will be used for for-profit tasks paid by data and processing providers using DAIA member project Tokens
2. Some will be used for tasks paid by EquiTokens
3. Some will be used for tasks that are chosen by democratic vote among network members, via a one-member-one-vote process (incorporating liquid democracy)
The percentage of effort to be spent on 1 is to be determined in terms of the percentage of data/processing contributions that are specified in terms of DAIA member project Tokens rather than EquiTokens. The percentage of effort to be spent on 2 versus 3 is to be decided by voting according to a one-member-one-vote process.
BACK-END FEATURES:
To realize this process and vision, the key technology components required are:
• Appropriate user interfaces
• Smart contract coding for the EquiToken, and cryptocurrency wallet coding for EquiToken accounts, etc.
• A powerful decentralized AI layer, which can be provided by collaborative effort from developers associated with DAIA member projects
• A powerful decentralized data-storage layer, which can be provided by appropriate DAIA member projects
• A decentralized liquid-democracy framework to support voting and delegation of votes
• A decentralized reputation system, enabling proposers of tasks for the Global Brain Network to be effectively rated by the community (so these ratings may be taken into account in voting). This requires use of machine learning to militate against “reputation fraud.”
• Highly refined security and anti-fraud mechanisms, leveraging AI from DAIA member projects as needed
• Inter-operation between the networks associated with the multiple DAIA projects operating on the back end of the system — in terms of identity and reputation management, token and data exchange, security and other functions
With the exception of the user interface, all of these tools are being developed within various DAIA member projects already, although not leveraged for this particular purpose. The conclusion is that no extremely major back-end technology development will be needed for the Global Brain Network. Although, the development of the EquiToken framework and user interfaces will not be trivial and will require attention from individuals with strong UI/UX and blockchain and computer security development experience respectively.
VERY GOOD NEWS:
This seems tractable to build within reasonable funds and time, and appears to have strong potential to have a very major positive impact on the future of humanity and other sentient on the planet.
CHALLENGES:
1. Getting adoption will be the main challenge. This will take a major, sustained PR push. The hope is that frustration with big tech companies and their data/AI oligopoly will drive adoption of this alternative platform, if it’s made easy and exciting enough to use.
2. Some of the key back-end technologies involved are not yet mature and are only recently launched in beta form or alpha form
3. For the next few years, fully decentralized processing and storage will be significantly less efficient than centralized alternatives. On the other hand this inefficiency may be in some regards compensated by unique privacy related features or clever AI algorithmic supplied by DAIA member organizations.
In short: Incredible transformative potential based on moderate but not humongous incremental technology effort; with some challenges regarding gaining adoption and accelerating progress in the face of dependence on early-stage back-end technologies.
How can you get involved?
The vision of DAIA is to foster a world where AI technologies and associated data are made open with decentralized, democratic control for the benefit of all sentient beings.
The immense potential of AI means that it can either increase the inequalities of our societies or liberate us from numerous sufferings. We believe the best way forward is to come together and work practically toward creating a better future. We see a massive potential for evolution in the established centralized corporations. We believe tech giants can contribute immensely toward making the vision of DAIA a reality.
DAIA welcomes the participation of those corporations that are sincere about their aim and goal of democratizing AI. The open access networks that have come together to form DAIA are the enabling layer for such a democratization process.
To learn more about us and inquire for memberships, please contact us at team@daia.foundation.
