The Twisted Road and AI

The book also a story about AI.

The book is set in 2294, a time when many of the disruptions and challenges posed by AI in the 21st century have been resolved. Sentient machine minds exist and function in a more-or-less benign way, providing vital knowledge and orchestrating many of the complex processes of a technological civilisation.

Aside from sentience, the main challenge that has been overcome is that of reliability. These are not the archaic AIs that hallucinate by blindly following patterns; instead, their pronouncements are guaranteed to be accurate and to conform to the actuality of the real world. These are the so-called canonical AIs, that rely on inputs from guaranteed reliable input hardware (data stores, cameras etc.) and that check what they say against reality.

These AIs function something like the security that exists today on web pages: we know we can trust pages because they have been digitally signed with a certificate created by a trusted issuer. Most AIs in The Twisted Road work that way: their authority ultimately comes from the root Earth AI, which guarantees that everything output is known to be true. Although it is also still possible to have AIs that don’t inherit from the reliable Earth AI in this way…

Canonical AIs are useful in many ways – one of which is in evidence-gathering for the forces of law and order. A canonical recording of witnessed events is always admissable in law, because it is guaranteed to be an accurate portrayal of what really happened.