Trustless Computing Initiative
& the future of AI
Special keynote by Rufo Guerreschi & panel discussion
Wednesday, November 9th 2016, 5-6:30pm
Link to the slides presented: http://bit.ly/2fs5WyQ
Symbolic Systems Program
Margaret Jacks Hall, Greenberg Room
Stanford, CA, 94305
ABSTRACT: Can radically-higher levels of assurance (i.e. trustworthiness) of deterministic IT systems – and related certification governance models – become a key factor to foster sustainable AI-driven economic development, as well as short- and long-term AI safety and value alignment?
If so, how can we prevent malevolent use and ensure constitutional lawful access?
Directions: (on Google Maps, building location on Campus Map)
Tickets: Event is free, but seat are limited, so please register on our Eventbrite page.
Program: 5-min intro by Todd Davies; 45-min presentation; 20-min discussant panel; 20-min Q&A.
We’ve been asked by the Associate Director of the Stanford University – Symbolic Systems Program (SSP), Todd Davies, to host a 1.5 hours presentation and discussion on our revolutionary Trustless Computing Initiative, its spin-off TRUSTLESS.AI, and their contribution to the economic and societal sustainability of critical AI systems. SSP is the largest post-graduate program at Stanford University dealing with AI, with over 90 professors and hundreds of students.
Uniquely to Trustless Computing, all and every software, hardware and processes that are critically involved in the ICT service provisioning or lifecycle – from CPU design to fabrication, from hosting room access to standard setting governance – are subject to (1) full verifiability w/out NDA and (2) extreme security review relative to complexity, and extremely resilient socio-technical oversight. It include innovative offline citizen-witness or citizen-jury processes applied to any server room access (CivicRoom) and to the most critical fabrication phases (CivicSite).
Trustless Computing will renounce completely to assumption of trust in anything or anyone, except in the intrinsic resistance of the organizational processes critically involved, as recognizable by moderately informed and educated citizens; while acknowledging its need to achieve an very high level of confidence to be resistant to attacks of tens of millions of dollars to compromise the life-cycle or supply-chain by actors with very low liability and high access to plausible deniability.
We are pursuing such goals through four main line of activities: (1) 4-14M€ research proposals to EU R&D public funds through the Trustless Computing Consortium; (2) a global events series, Free and Safe in Cyberspace with amazing speakers; (3) the drafting of a Proposal and Manifesto of Trustless Computing Certification Body (late draft gdoc); and, most recently, (4) via a startup spin-off just moved to Silicon Valley: Trustless.ai.
Once market-proven as leader in IT assurance – subsequent an initial go-to-market for mobile human communication – Trustless Computing technologies, ecosystems and certification governance processes will be extended to include ultra-high levels of availability — in addition to confidentiality and integrity – and then Trustless Computing aims to exceed ten-fold the trustworthiness of the most critical *deterministic* sub-systems (i.e. value systems, security monitoring, firmware upgrade, etc.) of mass-market safety and privacy critical autonomous systems, including drones, self-driving cars, robots, digital assistants, interactive toys.
The Initiative ultimate and sole goal is to substantially contribute to short- and long-term artificial intelligence safety, accountability and human values alignment.
- Rufo Guerreschi, Exec. Dir. of the non-profitTrustless Computing Initiative, and CEO of the spin-off startup Trustless.ai. (More on Linkedin or CV PDF)
Since 2013, as Exec. Dir. of the Open Media Cluster, Rufo has conceived and lead the Trustless Computing Initiative, Consortium and Manifesto, aggregating global partners and advisors with globally-rare or unique expertises, and open IP assets, in high or ultra-high assurance IT, spanning the entire IT lifecycle. Since 2015, he conceived, launched and coordinated the Free and Safe in Cyberspace event series (Brussels, Brazil, New York, Rome), which has attracted the leading EU and US experts and governmental institution in very focused discussions to build consensus on new standards for ultra-high assurance IT in critical market sub-domains, inspired by Trustless Computing concepts. Since May 2016, together with key advisors and partners of the Consortium have created the spin-off startup TRUSTLESS.AI as the main strategy to reach the goals of the Initiative. From 2008 till 2014 he was CEO and main shareholder of the Open Media Park, EU largest planned IT/media tech park, centered on open Web and cybersecurity, whose market valuation he brought from 3€ to 21M€. From 2002-2006, he founded and lead an open-source web solutions company, Participatory Technologies – based in Rome, Pune and Bangalore – that built e-participation solutions for social and political organizations, that sold in 3 continents. From 2001-2002, he lead EU sales of J2ME provisioning system for 4thpass (now Motorola) including a sale of 10M€+ to Telefonica.
Todd Davies, Associate Director of the Symbolic Systems Program at Stanford University. (Full profile at https://web.stanford.edu/~davies.) Todd teaches and does research on deliberation, collaboration technologies, social decisions, information policy, and collective behavior.
- Andrew Critch, Senior Fellow at MIRI (Machine Intelligence Research Institute), one of the leading long-term AI R&D center in the World. Andrew earned his PhD in mathematics at UC Berkeley studying applications of algebraic geometry to machine learning models. He co-founded the Center for Applied Rationality and SPARC. He left a position as an algorithmic stock trader at Jane Street Capital to join MIRI in September 2015. His current research interests include logical uncertainty, open source game theory, and avoiding arms race dynamics between nations and companies in AI development.