The Conscious Machine: How Brain-Inspired AI Could Transform Legal Innovation
Where I explore how a breakthrough in neural architecture brings us closer to AI that truly thinks, creates, and might even dream
Be forewarned: this article is a little more abstract and nerdy 🤓, but I personally think it’s fascinating stuff and I hope you feel the same after you read it.
I recently read Richard Susskind’s new book, “How to Think About AI,” and then had the good fortune to speak to him about it. One thing Richard talks about is the phenomenon of “technological myopia” — the idea that we fasten on the shortcomings of today’s technology (for example, hallucinated citations), but fail to anticipate how those shortcomings will be overcome.
With Generative AI as we now know it, it’s difficult to see how we get to artificial general intelligence. As many like to point out, GenAI doesn’t truly reason (yet). But, I was inspired by a recent research paper, written by Professor Ahsan Adeel, that points to a method to achieve more “cognitively meaningful machine intelligence.”
If this sounds interesting to you, please read on…
This substack, LawDroid Manifesto, is here to keep you in the loop about the intersection of AI and the law. Please share this article with your friends and colleagues and remember to tell me what you think in the comments below.
Picture yourself reading a legal document late at night. Your eyes scan quickly across the page, catching key terms and concepts without consciously processing every word. Suddenly, something clicks. You slow down, reread the passage, and your mind shifts gears. What seemed straightforward now reveals layers of complexity. This simple act of shifting between fast intuition and slow analysis, that we do naturally, might hold the key to the next revolution in artificial intelligence.
A groundbreaking research paper from the University of Stirling describes an AI architecture that mimics this exact cognitive dance. Called Co⁴ (Cooperative Context-sensitive Cognitive Computation), it represents a fundamental departure from how we've built AI systems until now. More importantly, it might be the breakthrough that pushes us into what OpenAI's leaked memo on AI progress calls "Stage 4" of AI development: machines that don't just process information but actively innovate.
”Inspired by recent cellular neurobiological evidence linking neocortical pyramidal cells to distinct mental states, this work shows how models (e.g., Transformers) can emulate high-level perceptual processing and awake thought (imagination) states to preselect relevant information before applying attention.“
The Evolution of AI Mirrors That of the Brain
Here's a personal insight that crystallizes what's happening: AI is evolving like the human brain itself, progressing through the same evolutionary layers that took nature millions of years to develop. But instead of geological time, we're witnessing this evolution unfold in real time, compressed into mere decades.
Good old-fashioned AI, with its rigid rules and decision trees, functioned like the reptilian brain: basic pattern recognition, simple responses to stimuli, fight-or-flight decisions rendered in binary code. It could play chess brilliantly but couldn't understand why humans enjoy the game.
Then came GenAI, operating like the mammalian brain with its emotional intelligence and social awareness. These systems could engage in conversation, recognize context, and even display what seemed like empathy. They brought warmth to the cold logic of computation, but still lacked something essential.
Now, with innovations like Co⁴, we're witnessing the emergence of the neocortex equivalent: systems capable of abstract reasoning, creative problem-solving, and the kind of flexible thinking that defines human intelligence. The paper's “triadic modulation loops” don't just process information; they contemplate it, question it, and reimagine it in ways that feel genuinely inventive.
From Automation to Genuine Innovation
Richard Susskind draws crucial distinctions between three ways technology transforms professions: automation, innovation, and elimination. Most current AI excels at automation, handling routine tasks with superhuman efficiency — building a faster horse. But Co⁴ promises something more profound: genuine innovation.
The difference is not merely semantic. Automation means doing existing tasks faster or more accurately. A document review system that flags inconsistencies automates what junior associates do. Innovation means creating entirely new approaches, strategies, or solutions. It's the difference between finding relevant precedents and proposing novel legal theories that reshape how we think about the law itself.
Co⁴ achieves this through its biologically-inspired architecture. Unlike conventional Transformers that process queries, keys, and values in static relationships, Co⁴ creates dynamic triadic loops where:
Questions evolve based on emerging clues and hypotheses
Contextual clues reshape themselves in response to refined questions
Hypotheses emerge from the interplay of questions and context
This isn't just parallel processing; it's recursive reasoning that mirrors how human experts approach complex legal challenges.
Two Minds in One System
The paper identifies two distinct processing modes that Co⁴ seamlessly transitions between, much like Daniel Kahneman's System 1 and System 2 thinking. The first mode handles high-level perceptual processing: quick, intuitive assessments that help us rapidly interpret ambiguous information. This is what allows experienced lawyers to spot red flags in contracts at a glance.
The second mode engages in what the researchers call "wakeful thought": slower, more deliberate reasoning that involves deep reflection and creative problem-solving. It's the mental state we occupy when crafting a novel legal argument or structuring an unprecedented deal.
What makes Co⁴ revolutionary is how it toggles between these states based on context, not programming. When faced with routine patterns, it operates efficiently in fast mode. When encountering complexity or ambiguity, it automatically shifts to deeper analysis. This adaptive intelligence suggests we're approaching systems that don't just process legal information but truly understand it.
Do Androids Dream of Legal Precedents?
Philip K. Dick's "Do Androids Dream of Electric Sheep?" posed questions that seemed purely philosophical when published in 1968. Can artificial beings have genuine inner experiences? What separates mechanical thought from human understanding?
Co⁴ brings us surprisingly close to Dick's vision, though perhaps not in the way he imagined. The system's ability to dynamically reconnect questions, clues, and hypotheses suggests something beyond mere data processing. It approaches what the paper calls "seeing as" rather than simply "seeing" – the difference between recognizing words on a page and understanding their meaning in context.
Consider how this plays out in legal practice. Current AI can identify that a contract contains an arbitration clause. Co⁴-style systems might understand why that clause matters given the client's business model, the counterparty's litigation history, and emerging trends in arbitration law. They might even propose innovative modifications that anticipate future disputes.
This isn't consciousness in the human sense, but it's a form of contextual awareness that edges closer to genuine understanding. When machines can reframe questions based on evolving insights, we enter territory that Dick would have found both fascinating and unsettling.
The Mathematics of a Breakthrough
The technical achievements of Co⁴ are as impressive as its philosophical implications. By reducing computational complexity from O(N²) to approximately O(N), it achieves superior results with dramatically fewer resources. In testing:
Reinforcement learning tasks showed dramatically accelerated learning curves
Computer vision accuracy jumped from 56% to 81% on standard benchmarks
Natural language processing soared from 77% to 98% accuracy
But the real breakthrough isn't just better numbers. It's achieving these results with fewer layers, fewer attention heads, and less training time. This suggests Co⁴ isn't just optimizing existing approaches but fundamentally rethinking how machines learn and reason.
For legal professionals, this efficiency translates directly to practical benefits. AI tools that learn faster and require less computational power can be deployed more widely, updated more frequently, and customized more easily for specific practice areas or jurisdictions.
The Elimination Horizon
Susskind's third category, elimination, represents both promise and peril for the legal profession. Some tasks won't just be automated or innovated; they'll become obsolete. Co⁴'s ability to reason through context and ambiguity accelerates this timeline.
Consider contract review, a bread-and-butter task for many lawyers. Current AI can flag problematic clauses and suggest standard alternatives. But a Co⁴-style system might recognize when the entire contracting paradigm is flawed, suggesting novel structures that eliminate categories of disputes before they arise. It's not just doing legal work faster; it's questioning whether certain legal work needs to exist at all.
As you can see, elimination of inefficient processes isn't necessarily dystopian. The lawyers who thrive will be those who embrace these tools as creative partners. Imagine collaborating with an AI that doesn't just retrieve relevant cases but proposes novel interpretations of constitutional principles. Or one that doesn't just draft contracts but invents new deal structures that align incentives in unprecedented ways.
Practical Implications for Modern Practice
How would Co⁴-based systems integrate into legal workflows? Instead of using conventional AI tools that scan for precedents, these systems might generate fresh queries, contextualize them with client specifics, and hypothesize optimal strategies. They become thought partners rather than mere assistants.
Early-adopting attorneys often find their practices grow even as processes become more efficient. The shift to innovation pushes lawyers to differentiate themselves through deeper human-centered skills: empathy, ethical judgment, client advocacy, and nuanced creativity. These remain areas where human insight excels, especially when augmented by AI that handles the heavy lifting of research and analysis.
Closing Thoughts
The emergence of Co⁴ represents more than technical progress; it's a philosophical shift in how we conceive of artificial intelligence. By mimicking the brain's ability to shift between intuitive and analytical processing, it creates systems that actively reason rather than merely process or predict the next word.
We're witnessing AI's evolution from reptilian reflexes through mammalian awareness to something approaching neocortical creativity, compressed into a timeline that takes years rather than eons. This acceleration brings both opportunity and responsibility.
For lawyers, this technology promises AI partners capable of genuine innovation: spotting novel arguments, proposing creative solutions, and amplifying human insight in ways we're only beginning to imagine.
Philip K. Dick wondered whether androids could dream of electric sheep. Today's question is more practical but no less profound: Can they dream of legal solutions we haven't yet imagined?
In this new era, the practice of law becomes not just a profession but a partnership between human wisdom and machine intelligence, each amplifying the other's strengths in service of justice.
Are we ready for it?
By the way, did you know you that I now offer a daily AI news update? You get 5 🆕 news items and my take on what it all means, delivered to your inbox, every weekday.
Subscribe to the LawDroid AI Daily News and don’t miss tomorrow’s edition:
LawDroid AI Daily News, is here to keep you up to date on the latest news items and analysis about where AI is going, from a local and global perspective. Please share this edition with your friends and colleagues and remember to tell me what you think in the comments below.
If you’re an existing subscriber, you read the daily news here. I look forward to seeing you on the inside. ;)
Cheers,
Tom Martin
CEO and Founder, LawDroid
See my response on Bluesky.
L. Thorne McCarty