What if the connection between human intelligence and artificial intelligence is not simply computing power, but chemistry?
A new paper from researchers with the Hebrew University of Jerusalem’s Department of Chemistry suggests that understanding memory may require scientists to look beyond neurons and electrical signals to the brain’s chemical makeup and environment.
Their recent study explores how interactions among neurons, astrocytes, and molecules surrounding brain cells could support the physical processes underlying memory, emotion, and thought.
The idea springs from something people are all familiar with: a simple song that suddenly brings back a childhood memory, a familiar smell that reminds someone of a place or person, or even a frightening experience that sparks a negative emotional or physical reaction.
“Biological creatures do not simply store information; they experience memory together with emotional and physiological states,” says researcher Dr. Gerard Marx of MX Biotech Ltd, one of the study’s co-authors.
“A traumatic memory, for example, is not just data. It can involve fear, bodily reaction, chemical signaling, and a felt state. At present, there is no formula or algorithm that encodes an emotive state in AI equivalent to the lived experience of biological organisms,” Marx adds.
One difference between computers and the human brain is that while computers can store and retrieve information, biological memory appears to involve something far more complicated and intricate.
Chemical memory
Marx, along with his colleague Prof. Chaim Gilon, draw on a theory they call the Tripartite Mechanism of Memory. The idea is that memory develops through interactions involving three components: brain cells, particularly neurons and astrocytes; the molecular environment surrounding those cells; and chemical substances, including metal ions, neurotransmitters, and gliotransmitters.
From this perspective, the brain may not encode and retrieve information through electrical activity alone. Chemical interactions could also play an important role in how information is physically written and read.
Astrocytes may also contribute by interacting with neurons and the brain’s chemical environment to help process and store information. The researchers suggest this pathway could extend beyond memory and potentially help explain aspects of thought itself.
In an email to The Debrief, Marx says this distinction becomes particularly important when comparing biological intelligence with AI.
“Understanding what makes biological mentality different from machine intelligence is scientifically essential,” Marx said.
He argues that while AI systems can process information, recognize patterns, and imitate aspects of human reasoning, biological intelligence appears to involve something more: “living tissue, chemical signaling, memory, emotion, and bodily states.”
“Our work attempts to clarify this distinction by proposing a biochemical mechanism for memory,” Marx explains. The model, he says, suggests that memory depends on interactions among neurons, astrocytes, the extracellular matrix, and chemical “dopants” such as metal ions, neurotransmitters, and gliotransmitters.
“If memory is not only electrical or computational, but also biochemical and emotive, then biological intelligence may be fundamentally different from machine intelligence, even when both can produce intelligent behavior,” Marx says.
Is memory more than information?
One obstacle to understanding memory is that memories do not exist simply as objective pieces of information. Looking more deeply at the physical nature of memory, the researchers argue that any comprehensive explanation must also account for its emotional dimensions.
This argument raises another question: Could the same pathways that help the brain store and retrieve information also contribute to thought, intelligence, and consciousness?
“Our research does not yet provide such a test, but it points toward what such a distinction might require,” Marx told The Debrief. “A machine may become extremely good at imitating conscious behavior, including language, reasoning, memory-like responses, and even expressions of emotion. But imitation is not the same as biological mentality.”
“If consciousness depends on the biochemical mechanisms that connect memory, emotion, and bodily life, then a purely computational system may only be able to simulate the outward signs of consciousness but without possessing consciousness itself,” he adds. “In that sense, our work may help frame future scientific criteria for distinguishing biological mentality from artificial imitation: not by asking only whether a system behaves intelligently, but by asking whether it possesses the biological and chemical mechanisms that generate emotive states and judgment based on that.”
What about AI?
As AI systems rapidly become more capable, questions about intelligence and consciousness are becoming increasingly difficult to ignore.
Today’s AI systems can store and process information, identify patterns, solve complex mathematical and informational problems, and generate responses that can appear remarkably intelligent. However, there is currently no established evidence that AI systems experience information or the world in the way humans do.
Marx and Gilon’s work challenges the assumption that simply increasing an AI system’s computational complexity would inevitably produce consciousness. The researchers do not claim to have solved the consciousness enigma. Instead, they propose a biochemical framework for investigating how physical processes in the brain might connect memory with emotion, cognition, and thought.
Elizabeth (Liz) Ngonzi, of the American Society for Artificial Intelligence (ASFAI), offered a different perspective in an email to The Debrief, noting that she doesn’t see “a fundamental separation between humans, living nature, and technology.”
“Everything is energy and information,” Ngonzi said. “Just as humans learn by absorbing cultural knowledge, fine-tuning through experience, and adapting through feedback, AI models are trained on, fine-tuned by, and animated through our collective human knowledge and interaction.”
“AI is not a separate, alien entity—it is a technological mirror of how collective intelligence evolves.”
“Trying to build ‘conscious AI’ out of silicon confuses calculation with living awareness,” Ngonzi added. “It’s a category error. Instead of chasing the illusion of machine consciousness, our real work in AI governance is Co-Intelligence, using AI as a non-conscious amplifier of human capability, while keeping human wisdom, embodiment, and ethical stewardship firmly at the center.”
AI Whistleblowers
The discussion comes amid broader concerns about the future development of increasingly capable AI systems. Former Anthropic AI researcher Jacob Coxon recently left the company and the industry and publicly warned that he believes advanced AI could pose an existential threat to humanity by 2030.
“The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible—but I hear the same people express fear privately. No other human activity poses this level of danger,” Coxon recently told The Guardian.
Ngonzi argues that biological life and technology should not be viewed as competitors for consciousness, but as interconnected expressions of energy and information. Rather than attempting to recreate human consciousness in machines, she believes AI should be developed as a tool for “Co-Intelligence,” amplifying human capabilities and relational intelligence.
“When we stop trying to force silicon to recreate organic awareness and instead design AI to amplify human relational intelligence, we move from the distraction of artificial consciousness to the true promise of Co-Intelligence,” Ngonzi told The Debrief.
But the debate over AI risk also raises a more fundamental philosophical question: What is intelligence?
Marx and Gilon’s work could ultimately help researchers investigating the relationship between AI and consciousness approach that question from a different direction. If human thought emerges from more than electrical signals—if chemistry, cells, and embodied experience are also part of the equation—then understanding consciousness may require us to look not only at how information is computed, but at what kind of system is doing the computing.
Chrissy Newton is a PR professional and founder of VOCAB Communications. She hosts the Rebelliously Curious podcast, which can be found on The Debrief’s YouTube Channel. Follow her on X: @ChrissyNewton and at chrissynewton.com.