Making Morally 
Machines
Table of Contents
Demonstrations
- Watch here the robots that pass one of Luciano Floridi’s harder-than-the-familiar-"Wise Man"-puzzle, and in so doing display what’s needed to be a morally self-aware robots.
- Automated-reasoning technology enables this AI system to use the Doctrine of Double Effect to solve the classic “trolley problem.”
- We all know what “weakness of the will” is (or — to use the fancy term of the philosophers — akrasia): You ought to do something, know that you ought to do it, and indeed want to do it … but then, alas, at the moment of truth, you don’t meet your obligation. Watch as this robot emerges victorious over the urge to exact revenge.
Exploring the Underlying Logic & AI for Making Morallly X Machines
Those who want to explore the cognitive calculi and AI technology that enable us to engineer morally correct machines and thereby surmount the PAID problem are invited to do so. Such explorers can take one or more of three initial routes:
- use the HyperSlate® proof- and program-engineering environment from Motaline to learn and explore the cognitive and ethical calculi at the heart of our R&D;
- learn and explore the Naveen-Sundar-G-led “Shadow” family automated reasoners, starting with ShadowProver;
- learn and explore NSG’s automated planner, Spectra.
Links to these systems are given below.
Would-be explorers with slim background in formal logic & AI may wish to enroll in either Introduction to Formal Logic and AI or — for those who have taken a first course in formal logic and/or AI — Intermediate Formal Logic and AI. Both courses taught by Bringsjord introduce, teach, and use the trio of systems for AI and cognitive robotics mentioned in the previous paragraph, links to which are here:
Media (selected)
- In NHK World’s “AI and Ethics,” available here, Bringsjord starts by sharing the “PAID” problem.
- The impossibility of morally heroic robots, discussed in Bringsjord’s TedX Talk in Cyprus.