Sapience.org is a Thinktank: A Public Policy Research Institute based in London, UK
Sapience.org is a London-based think tank: a public policy research institute focused on the future of progressively intelligent computing and its disruptive effects. Our work on technological forecasting and artificial intelligence informs corporations and policy makers on disruptive computing technology and on effective methods of managing its risks.
We advise venture capital investors, blockchain start-ups, health organizations, policy makers, and charity groups on the disruptive impact of artificial intelligence and increasingly disruptive computing technology. We support strategic technological decisions requiring answers to questions such as: What will succeed the transformer architecture, and on what timescale? How will AI affect computer trading, now that almost all stocks are traded by algorithms? Which decisions may a board lawfully delegate to an autonomous system under the EU AI Act? How will driverless cars affect transportation? What evidence would show that an AI system does what it was specified to do?
Motivation
Public debate about artificial intelligence is no longer speculative. Large language models have put general-purpose systems into everyday use, AI safety has become an established research field, and on 2 August 2026 the European Union’s AI Act came into force — the first comprehensive statutory regime for the technology. Boards, regulators and governments must now make decisions that turn on what these systems are and what they can be shown to do. Steven Hawking and colleagues have appealed to the public after the release of the film Transcendence, urging against complacence about AI risks. That debate remains fuelled by confusion about what exactly is being claimed. Research we conducted in Singularity Hypotheses (Springer 2013) since 2009 shows that the debate is often ill-informed. In particular, the term ‘technological singularity’ describes two distinct and entirely different scenarios. And critiques that aim their objections at the plausibility of apocalyptic scenarios are often distracted by the singularity ‘meme’ — such as Skynet in the Terminator franchise — where in fact the literature is entirely focused on risks from indifferent, not malevolent, AI. The same failure now attaches to ‘alignment’, ‘agency’ and ‘general intelligence’: the argument moves faster than its definitions. You cannot govern what you have not defined.
