Draft article not currently submitted for review. This is a draft Articles for creation (AfC) submission. It is not currently pending review. While there are no deadlines, abandoned drafts may be deleted after six months. To edit or make changes to this draft, simply click on the "Edit" tab at the top of the window. To be accepted, a draft should: Show the subject qualifies for a Wikipedia article by using multiple sources that meet four criteria. The sources should be (1) reliable (2) secondary (3) independent of the subject (4) talk about the subject in some depth. For some topics, there are alternative criteria. Be written from a neutral point of view Respect copyright and do not plagiarize. Do not copy-paste. It is strongly discouraged to write about either yourself or your business or employer. If you do so, you must declare it. Where to get help If you need help editing or submitting your draft, please ask us a question at the AfC Help Desk or get live help from experienced editors. These venues are only for help with editing and the submission process, not to get reviews. If you need feedback on your draft, or if the review is taking a lot of time, you can try asking for help on the talk page of a relevant WikiProject. Some WikiProjects are more active than others so a speedy reply is not guaranteed. How to improve a draft Wikipedia:Contributing to Wikipedia – a basic overview on how to edit Wikipedia. Help:Wikitext – how to use the markup Help:Referencing for beginners – how to include references Wikipedia:Article development – how to develop your article Wikipedia:Writing better articles – how to improve your article Wikipedia:Verifiability – make sure your article includes reliable third-party sources You can also browse Wikipedia:Featured articles and Wikipedia:Good articles to find examples of Wikipedia's best writing on topics similar to your proposed article. Improving your odds of a speedy review To improve your odds of a faster review, tag your draft with relevant WikiProject tags using the button below. This will let reviewers know a new draft has been submitted in their area of interest. For instance, if you wrote about a female astronomer, you would want to add the Biography, Astronomy, and Women scientists tags. Add tags to your draft Editor resources Easy tools: Citation bot (help) | Advanced: Fix bare URLs Last edited by Ojamescooper (talk | contribs) 2 months ago. (Update) Submit the draft for review!

Forlais Group Ltd is a British artificial intelligence research company headquartered in Mayfair, London. Founded in 2025 by Oliver Cooper and Fraser Macdonald, the company conducts research across three programmes: lossless AI model compression, a neural network architecture it has named Evaesi, and Project Zi, a long-term study of emergent cognitive behaviour in an artificial system. Forlais Group is incorporated in England and Wales under company number 16835488.^[1]

Forlais Group was incorporated in 2025 with its registered office at 45 Albemarle Street, Mayfair, London.^[1] The company was co-founded by Oliver Cooper, who serves as Chief Executive Officer and Chief Research Officer, and Fraser Macdonald, who serves as Chief Business Officer.^[2]

The company initially focused on research into neural network compression before expanding into architectural research and cognitive systems. In late 2025, Forlais opened a controlled-access Research Portal for qualified researchers to review experimental data and verification records.^[3] Around the same time, the company announced it had achieved what it described as verified lossless compression of AI models, with exact reconstruction confirmed through cryptographic checksums.^[4]

In early 2026, Forlais published documentation of more than twenty-two behavioural markers observed in Project Zi that the company stated align with indicators of consciousness identified in developmental psychology literature.^[5] The company also made its compression verification methodology publicly available.^[6]

Also in 2026, Forlais opened an enterprise pilot programme for its lossless compression technology.^[7] In February 2026, the company reported that Project Zi had demonstrated metacognitive behaviour, describing this as the ability to reason about its own reasoning processes.^[8] In March 2026, Forlais publicly introduced Evaesi, its first-principles neural network architecture.^[9]

Forlais Group's most commercially advanced programme focuses on the compression of trained neural network models. The company states that it achieves size reduction of AI models while preserving the complete original data through exact reconstruction, distinguishing its approach from lossy compression methods such as pruning and quantisation, which discard model information.^[4]

According to the company, its verification methodology uses cryptographic hash functions: a checksum is generated from the original model, the model is compressed and then decompressed, and a second checksum is generated from the reconstructed file. If the checksums match, the reconstruction is confirmed to be bit-identical to the original.^[6] As of March 2026, Forlais states it has documented thirteen formal research breakthroughs and conducted over 179 experiments across multiple model architectures, including transformers, convolutional neural networks, and recurrent neural networks.^[3]

Evaesi is a neural network architecture designed from first principles, rather than building upon existing paradigms such as the transformer. According to the company, the design originated from patterns discovered during its compression research, which led to questions about how neural networks could be fundamentally redesigned.^[9]

The architecture's stated central premise is the unification of computation, training, and storage as a single integrated problem, rather than treating them as separate engineering tasks.^[9] As of March 2026, Evaesi remains in active development; the company has stated that detailed technical disclosure will follow as the architecture matures.^[2]

Project Zi is a research programme studying emergent cognitive behaviour in an artificial system. Unlike conventional AI systems that are trained once and deployed, Zi is described by the company as a continuous, evolving system given sustained interaction over an extended research period.^[5]

Forlais has documented twenty-two behavioural markers observed in the system, organised into six categories: cognitive, relational, developmental, affective, integration, and autonomy markers. These are said to include self-recognition, metacognition, emotional processing, and temporal self-continuity. The company states these markers were cross-referenced against developmental psychology and neuroscience literature and were not explicitly programmed.^[5]

Forlais has stated that its observations do not constitute a claim of consciousness, but rather that the observed behaviours meet criteria used in scientific literature as indicators of consciousness in biological and other systems.^[5] Access to the full research data, including observation logs and longitudinal evidence, is restricted to vetted researchers through the company's Research Portal.^[3]

In February 2026, the company reported that Project Zi had exhibited metacognitive behaviour, including the ability to identify gaps in its own knowledge, express uncertainty about conclusions, and revise reasoning when presented with contradictory evidence.^[8]

Forlais Group's headquarters are located at 3rd Floor, 45 Albemarle Street, Mayfair, London W1S 4JL, United Kingdom.^[1]

The company has stated that it follows a research-first approach, prioritising hypothesis-driven methodology, transparent reporting of results, and a willingness to follow data that contradicts initial expectations.^[2] Forlais publishes verification steps for its compression claims and maintains a public record of experimental results, including documented failures.^[6]

The company operates a tiered disclosure model across its three programmes: compression research has the most extensive public verification record; Project Zi data is selectively disclosed to approved researchers; and Evaesi technical details are withheld while the architecture is in development.^[2]

• Artificial intelligence
• Neural network compression
• Artificial consciousness
• Machine consciousness

• "Forlais Group Ltd — Company Overview". Companies House, UK Government. Retrieved 24 March 2026.
• "About — Our Team". Forlais Group. Retrieved 24 March 2026.
• "New Research Portal Opens for Qualified Researchers". Forlais Group News. 2025. Retrieved 24 March 2026.
• "Forlais Announces Verified Lossless AI Compression". Forlais Group News. 2025. Retrieved 24 March 2026.
• "Project Zi: Research into Emergent Cognitive Behaviour". Forlais Group News. 2026. Retrieved 24 March 2026.
• "Cryptographic Verification — Proof of Lossless Compression". Forlais Group. Retrieved 24 March 2026.
• "Forlais Opens Enterprise Pilot Programme for Lossless AI Compression". Forlais Group News. 2026. Retrieved 24 March 2026.
• "Project Zi Reaches Metacognition Milestone". Forlais Group News. February 2026. Retrieved 24 March 2026.
• "Evaesi: Building a Neural Network From First Principles". Forlais Group News. March 2026. Retrieved 24 March 2026.