SACRED IDEAS · AI & SOCIETY
AI, consciousness and the society we are making.
A machine can help us think, write and investigate. What follows when we begin to rely on its judgement? Sacred Ideas examines AI research and its societal consequences alongside the human questions of consciousness, inner experience and responsibility.
WHY THESE QUESTIONS BELONG TOGETHER
What does it mean for something to matter?
When a person turns to an AI companion in grief, its words may offer comfort. That tells us something about their effect on a person. It does not establish that the system feels concern, understands loss from within or experiences anything at all.
The distinction reaches into Sacred Ideas’ oldest questions. Love, grief and moral obligation have a felt dimension in human life. Explaining their physical basis does not make that experience disappear. But a machine’s ability to speak persuasively about them raises a further question: what would count as evidence of experience in an artificial system?
Our interest in AI grows from this shared territory, and from the consequences of introducing powerful systems into lives and institutions already struggling with meaning, trust and authority.
CONSCIOUSNESS & INNER EXPERIENCE
Keep three questions separate.
What can the model do?
A foundation model is a trained computational system. Its abilities to generate, analyse, reason or solve problems are questions of capability, to be assessed on evidence.
What can the system carry through?
An agentic system adds an operational architecture: tools, memory, goals, planning and permissions to act. Functional agency and autonomy concern how that system operates.
Is anything experienced from within?
Possible subjectivity is a different question. Intelligence, language, self-description and autonomous action do not by themselves establish that there is anything it is like to be the system.
We assume neither that sophisticated machines must become conscious nor that non-biological systems cannot. The question requires philosophical argument and scientific evidence. Human experience is the starting point of the inquiry; its artificial counterpart remains an open question.
DEVELOPING RESEARCH
When AI helps to build AI.
We are developing a documentary inquiry into AI participation in research, including research on AI itself. The work follows how systems generate proposals, run experiments, filter candidates and help evaluate results—and what humans are actually able to check.
A team may retain formal control of a workflow while depending heavily on machine-generated analysis. Being able to monitor its steps is not necessarily the same as being able to independently verify a claim or meaningfully challenge a recommendation. Under what conditions does useful assistance become deference to a system’s judgement?
That is a question to test, not a conclusion to announce. Automation can also improve verification. Our inquiry examines the availability of independent checks, the limits of human review and the authority institutions grant to outputs they accept. Forecasts of an “intelligence explosion” are not treated as observations.
SOCIETAL CONSEQUENCES
Who can challenge a machine-supported decision?
Questions about work, education, public information and democratic participation cannot wait for a verdict on machine consciousness. We examine how AI developments may widen access to knowledge and useful services, reshape employment, concentrate power or change whose judgement receives a hearing.
Consider a school. AI assistance can support a student’s developing judgement, or make it easier to bypass the work through which judgement grows. What matters is how the tool is used, what the institution rewards and whether students learn to question the answers they receive.
The same demand for accountability belongs in public life: people affected by a decision need meaningful ways to question its evidence and seek correction. Technical capability does not, on its own, confer legitimate authority. And a system can faithfully pursue a human objective that should never have been chosen.
THE INCANDESCENCE CONNECTION
What kind of human being does this system make more likely?
Incandescence names the inward dimension of nature disclosed through conscious life. In human life, it also asks how we cultivate attention, care, judgement and responsibility. This is a philosophical framework, not a scientific finding that the cosmos is conscious.
AI gives the institutional question new force. What habits of attention and reliance are we encouraging? What remains open to correction? Machines may help us pursue an end; the worthiness of that end remains a human question.
EVIDENCE & RESEARCH PRACTICE
Starting points in the literature.
We work from primary research and documented systems, distinguish reported results from interpretation, and test promising claims against their strongest objections. These selected sources are starting points; they do not establish our developing research conclusions.
- Butlin and colleagues, Consciousness in Artificial Intelligence (2023): a theory-informed approach to proposed indicators of AI consciousness.
- Yamada and colleagues, The AI Scientist-v2 (2025): a documented agentic research workflow, to be assessed within its demonstrated scope.
- Lindsey and colleagues, On the Biology of a Large Language Model (2025): circuit-tracing case studies and explicit limits on what their interpretability methods reveal.
New essays, talks and research findings will appear as the work earns them.
CONTINUE THE CONVERSATION
Follow the questions as they develop.
Receive Sacred Ideas essays and updates on religion, consciousness, AI research and public life.
Occasional emails. Unsubscribe at any time.