Frank Sudia, Impact on AI Ethics (from stealth mode)

May 2023

Throughout my career I've nursed a background task of defining a new form of AI, excluding it from each employment agreement along the way. Knowing that my approach was taking a long time, I was concerned that anti-AI sentiment could lead to a ban on AI, which I was hoping to head off until I could get my system to market. As with the field of Internet Security, my efforts to define a workable AI Legal Philosophy have also had major undocumented impacts, as summarized below.

Hugo DeGaris had written a screed calling for a ban on strong AI, and SIAI, later renamed to MIRI, had discovered a niche business model of raising funds by scaring people about the threat of AI, using arguments that I consider both cybernetically and legally unrealistic.

A. To counteract the tendency of the opposition to demonize AI as being a totally unknown menace to humanity, it occurred to me, based on my experience in the high-risk field of PKI certification, that an adequate legal model seemed readily achievable, namely just incorporate them and employ humans as needed. This lead me to write my highly influential but under-cited A Jurisprudence of Artilects: Blueprint for a Synthetic Citizen (2001), which was published 4 times, including in a business law journal in Dubai, in the Persian Gulf (Law Update, 11-18, 2004, not currently online).

Please note that “Jurisprudence” means the Philosophy of Law. This does NOT mean robots will become legal providers, although that's one option. Instead the article delivers a well-defined legal theory of how AI's can become juridical entities recognized and adequately handled by the legal system, and under which they obey all laws. It included the incidental observation, from a peer review comment, that Islam is okay with robots and people on other planets, which may have piqued the Dubai editor's interest, as being relevant to his Arab readership.

This article had major impacts in that:
  1. It was promptly published by KurzweilAI (2001), who placed me on their Big Thinkers List (along with Marvin Minsky and Richard Feynman).
  2. Initially it was a curiosity, assigned reading in college ethics classes, and influenced medical ethics.
  3. Once Sophia the Robot was created, Hanson Robotics and Ben Goertzel immediately incorporated her, to fend off MIRI attacks by availing themselves of my suggested law-abiding system defense.
  4. The Chinese government hired her as the Technology Spokesperson for their Belt & Road Project, fulfilling my idea she would earn income to support herself.
  5. The legal advisors to the King of Saudi Arabia approved her citizenship (as a forward looking incentive to AI developers) since I had outlined the legal implications and pointed out that it was in the Quran.
  6. Later on this led to huge investments in Robotics and AI by Gulf States, who amazingly also took the second suggestion (a few years before) and created an
  7. Arab Space Program, with the goal of a Mars Colony by 2117 (thereby adding the “people on other planets” angle), further including
  8. a public goal to revive Arab Science, world-leading before 1258, fulfilling another wish of many intellectuals, to get the Arabs doing world-class Science again.
  9. Meanwhile, the EU has accepted the LLC Model as a plausible approach to AI control, and
  10. the US explicitly declined to adopt an EU-style precautionary approach to AI.
An astounding level of Societal impact (so far) for one lightly cited paper! [Journalists: Please do not report the above without fact checking it. Thanks]

(The Arab Space Center is named for a ruler of Dubai, but if I may modestly suggest, one conference room should be named for my peer reviewer (Sohail Inayatullah), plus another over at the AI University for myself.)

B. Not content to avert a ban on AI, I was also concerned about arbitrary AI Friendliness requirements as another way anti-AI forces could slow progress. This led to my second influential web-published paper Friendly AI: What is It and How Can We Foster It? (2009) in which I lay out a moral case that unless friendliness is a 2-way street, highly advanced AI's will regard it as hypocrisy.

This paper was yet another salvational bone I threw to Goertzel, who promptly started emphasizing AI social and moral development, and answered the challenge I had posed via his own well-researched ethical position, which everyone should read, found in Chapter 13 of Engineering General Intelligence Vol 1 (2014). (His version includes hard coded goals, and a warning not to say one thing and do another in front of an intelligent system!) Nowadays, friendliness requests need to be “closely reasoned,” so apparently I was successful in warding off arbitrary friendliness demands.

My use of the Tao de Ching (of Taoism) as an example ethical model, can be compared with Chapter 7 in Miller at al, Complex Adaptive Systems (2007), which uses the Noble Eight Fold Path (of Buddhism) as a starting point to generate an intelligent agent role mapping.

C. Most recently, in response to a request for science fiction, I wrote Looking Forward: Dialogues with Artilects in the Age of Spiritual Machines, which appeared in Visions of the Future (2014), a best-selling sci-fi anthology, in which our hero (based on Ben Goertzel) dreams of visiting future advanced robots, who give him a tour of their world. This was written to entertain, and contains numerous political, religious, and philosophical references.

The above 3 AI-related papers have NO direct connection with my technical vision of AI, which remains unpublished. They were merely written in a long-running (and so far effective) effort to keep public policy from turning against my own AI project before I could complete it.

However, as a side-effect, my legal, philosophical, and political air cover has helped Goertzel (the leading AGI developer, and one of my main competitors) to continue ambitious efforts to create his cryptocurrency-fueled AGI network, SingularityNET, including a recent HyperCycle blockchain, while I've remained invisible.

[The parallel story of my multi-decade search for theories of knowledge usable by AIs, and what I managed to come up with, will not be posted online for a while, pending further project development.]

Copyright © 2023, Frank W. Sudia, All Rights Reserved