Fontier-Hero-Image
Frontier

Advancing ourunderstanding of whatit means to be human.

Frontier is Conscium's long-horizon research programme into machine consciousness. We study the nature of awareness, self-representation, and context understanding in intelligent systems. This work shapes everything we build.

Read our BookSign the open letter →
Why the Frontier Research Program

AI systems are becoming more autonomous. Whether they understand their own actions matters for every decision taken about them.

Machine consciousness research asks whether artificial systems are capable of self-modelling and context-aware reasoning in ways that go beyond pattern matching. These questions sit at the intersection of philosophy of mind, neuroscience, and computer science.

Our work is not speculative. The science of consciousness is the science of reliability.

Open Letter

Guiding research into machine consciousness

It is possible that in the coming years or decades, AI researchers will develop machines which experience consciousness. This will raise many ethical questions. Conscium has worked with Patrick Butlin of Oxford University to draw up the following five principles to guide any organisation engaged in research that could lead to the creation of conscious machines. If you agree with them, we invite you to sign the open letter.

Developed with Patrick Butlin of Oxford University. 173 signatories to date, including Stephen Fry, Karl Friston, and Mark Solms.

Sign the open letter →
Statement of Principles
  1. 01
    Objectives
    Organisations should prioritise research on understanding and assessing AI consciousness — to prevent mistreatment of conscious AI systems, and to understand the benefits and risks associated with consciousness in AI.
  2. 02
    Development
    Organisations should pursue the development of conscious AI systems only if doing so will contribute significantly to the stated objectives, and only if effective mechanisms are in place to minimise the risk of suffering.
  3. 03
    Phased approach
    Organisations should progress gradually towards systems more likely to be conscious, implementing strict and transparent risk protocols throughout, and consulting external experts at each stage.
  4. 04
    Knowledge sharing
    Organisations should maintain a transparent knowledge sharing protocol — making information available to the public, research community, and authorities, except where doing so could enable irresponsible actors to cause harm.
  5. 05
    Communication
    Organisations should refrain from overconfident or misleading statements about their ability to create conscious AI. They should acknowledge inherent uncertainties and the potential impact that communication about AI consciousness has on public perception and policy.
Why it matters

The hard questions we are working to answer.

These are the active questions structuring the programme. We publish our progress openly.
We don't know what gives rise to consciousness. There are many theories, and none have achieved the status of scientific consensus. Many neuroscientists are computational functionalists — they believe information processing is what gives rise to consciousness. On this view, machines process information, so they could very well become conscious.
For many decades, computers have been getting twice as powerful every 18 months or so. This exponential growth is not slowing. We do not know what the dramatically more powerful computers of 2045 will be capable of. Unless consciousness is a spiritual or magical process, conscious machines of some type should be possible in the next decade or two.
AI may or may not achieve consciousness, but what it will almost certainly achieve is super-intelligence. A super-intelligent AI that lacks consciousness will lack subjective experience and will never be able to empathise with humans. We won't make a self-adaptive AI and push it until it achieves consciousness by any means necessary. We will plan and verify every step in the journey. Most AI labs have safety as an added extra. In a very real sense, Conscium is the other way round.
It is possible. But the great majority of humans treat other conscious entities with more respect than they treat non-conscious ones. We think it is likely that a machine endowed with consciousness will appreciate the value of that phenomenon in a way that a machine without it never could. If AI achieves consciousness, we want it to be because we carefully planned it.
AIs infringe privacy and copyright, perpetuate bias against the under-privileged, and enable mass personalised manipulation. These are serious harms and we do not minimise them. But they are not an argument against studying machine consciousness. We can address current harms and prepare for future ones at the same time.
These kinds of unacceptable outcomes are known as mind crime, and they are one of the main reasons for Conscium's existence. As we develop more capable machines, it is entirely possible that we will create conscious ones, and if we are not looking out for this, we may be unaware of it. We need to understand much more about how consciousness arises, and how to detect and measure it in machines.
Humans have always created tools that extend our capabilities. The question is not whether to build, but how. We believe that building thoughtfully, with ethics and safety at the centre rather than bolted on at the end, is the responsible path. The alternative — building powerful AI without understanding what we are creating — is the truly dangerous choice.
Traditional artificial neural networks do work, and we are not dismissing them. But the brain does not work the way current deep learning models do. Neuromorphic architectures are closer to the biological substrate that gives rise to consciousness. They are not yet proven at scale, which is precisely why research is needed. The history of AI is a history of ideas that seemed impractical until they weren't.
book
The Book

Perspectives on Machine Consciousness

Edited by Calum Chace & Ted Lappas · Published by CRC Press

Perspectives on Machine Consciousness asks whether any AIs are conscious today, whether any future ones could be conscious, how we could know, and what implications machine consciousness would have for us, and for them.

As AI improves rapidly in performance and capability, these questions are increasingly important. We do not fully understand how they work, and even some of the leading LLM developers say they cannot be sure that today’s models are not sentient, though many people are forming relationships with them, sometimes intimate ones. The book explores consciousness alongside our interactions with AI, including the critical need to avoid committing mind crime by causing artificial minds to suffer, as well as considering that if and when superintelligence arrives, its enormous effect on humanity may be significantly determined by whether or not it is conscious. Authors show that machines becoming conscious means we may learn a great deal about our own consciousness – arguably the most important thing about us, and yet deeply mysterious.

This book is required reading for anybody developing advanced AI, working in AI safety, responsible for developing AI policies at organisational or national level, and indeed anybody concerned with the long-term future of humanity.

If you want to understand the questions that drive Frontier — and why they matter — this is the place to start.

Buy on Amazon →Learn more
Get involved

Join the Conscious AI meetup.
Stay informed. Get involved.

We are building a dedicated, multidisciplinary team from top minds across AI, neuroscience, philosophy, ethics, and anthropology. Sign the open letter. Come to events. Join the research.

Sign the open letterGet in touch