The threats and danger of unsafe AI are known to us. The unknown is who to trust and how to work together with those that you do trust to overcome and surpass to add to humanity's peaceful and sustainable future.
AGI has been a black box for us in this regards up to now. I propose the co-development of a new technology that would involve accountability and transparency in the collaborative examination of levels of evidence for what works best in a matrix of interconnectedness, overcoming artificial boundaries together, including:
- Ego and local positions of power
- differing interpretations of 'ethical', 'acceptable', 'lawful', etc.
- layperson access to algorithms, including AI and applied epidemiology
- differences in vocabulary and languages in humans, systems and fields of knowledge
- gender-based, cultural, religious and other differences leading to bias and opposing opinions
- field and depth of knowledge and expertise in skills
- different motives and priorities, including power, material or spiritual driven
- varying access to necessary resources for local and shared success.
The basis of the proposed AGI is an umbrella primitive ontology as a human understandable Formal Language that is co-developed in algorithmic strings, computatively guiding and optimising a shared commitment to develop and maintain shared problem-solving best practice. Collaborators continue to safeguard and improve the AGI and its tethered AI through a holistic and collective process of AGI-supported learning, education, critical thinking, consensus decision-making, joint action and continuous improvement for humanity's good, as part of the transparent peer-reviewed consensus process.
Inherent in this technology are continuously improving AI and AGI mechanisms as enablers for personal and shared growth supporting the development of 'allied' global citizens where differences in beliefs, appearances, knowledge and actions gradually become apparent, celebrated for ‘goodness’, and aligned through scientific-level, shared peer-review at a global level of what has historically led and what currently leads to the interconnected achievement of individual, group, institutional, national and global success vs. failures; satisfaction vs. dissatisfaction; good health vs. preventable morbidity and mortality; etc.
With this approach for a new internet platform for collaboration for the 'good', its continuously improving AI becomes the product and servant of the cohesion of and commitment to the principles underpinning its AGI, strengthening and guiding and becoming stronger over time against its opposition and attackers with the support of dedicated expert collaborators. The AGI also acts as shared algorithmic 'language' for legal smart contracts in a matrix of blockchains, depicting the growth of strong alliances for real-world problem-solving, its collaborative decision-making processes and shared knowledge, including the permissioned flow of resources and continuing peer-review evaluations with concomitant growth of shared wisdom. In association, its ‘current’ best practice algorithmic strings for various functions and tasks of ‘Safe’ AI are also safeguarded through blockchain technology, and continuously peer-reviewed and improved at a shared wisdom level, best countering threats and embracing opportunities while serving as a communication, education, research, and consensus development method at a human-understandable question and answer level in a shared model of the world. Its increasingly ‘Safe’ AI supports the identification of statistical outliers, both of opportunities and threats, to be examined and correctly identified for priority action, including the early identification of ‘unsafe’ collaborators attempting to corrupt the shared wisdom, and of true innovators with ‘first’ break-throughs.
Such a ‘good’ human in the loop ‘safe’ AGI/AI platform for concurrent collaboration, learning, research, work, risk and resource management can potentially bring about a paradigm shift to mankind's safety and quality of life for the long-term. Such an endeavour can only be achieved with well-intentioned expert collaborators globally carefully and deliberately examining and building on its theory, developing its technology and establishing its strong foundations, to best empower and safeguard those who wish to help bring it about.
Yours sincerely, Dr. Lisa New
MBChB, MBA (LBS)
Independent Researcher for Achievement of Human Rights assisted by a Fusion of Emerging Technologies
- algorithms blockchain machine learning Ethics Justice human rights deep learning AGI collaboration safe AI new internet platform primitive ontology formal language layperson consensus development decision support peer-review scientific method lawful organic learning smart contracts global citizens