Sterile verbosity, circular paraphrases, fabricated bibliographic references: such were the unfortunate specialties of novice artificial intelligences; a trifecta of shortcomings that reflected their nature more clearly than their intelligence. Behind a promising name hid a much murkier reality. Far from the clarity of spirit it claimed to embody, AI struggled to rise above simple mimicry. But the shift was as rapid as it was radical. The field has evolved swiftly: each new iteration emerges more sharpened, designed to iron out its own flaws, correcting its own blunder. 

Rumor has it: AI is merely a copy of the human brain's functioning. A key function of the human brain lies in the ability of its neurons, structural and functional units, to memorize actions through experience. Learning to drive, particularly, learning to change gears in a manual car, is a perfect illustration. Once driving is mastered, changing gears becomes more natural, an almost automatic gesture. 

At first, gear-changing is not instinctive. A beginner's motor cortex must activate different muscle combinations, leading to clumsy and incorrect movements. This process of trial and error repeats until the correct, successful sequence is achieved.

When this happens, the brain's reward system kicks in and generates a feeling of satisfaction that motivates repetition. Through practice, synapses, functional junctions between two neurons, grow stronger making them more receptive. The more a neural network is used, the stronger and more precise it becomes.

This same logic underpins artificial neural networks. These systems are structured in layers, with each neuron connected to all the neurons in the next layer. Initially, the connections have uniform weights, so the network's first attempts are often wrong. Just like for a student driver, feedback comes into play. The algorithm gradually adjusts the weights of the connections, and with each iteration, the network fine-tunes its parameters until it generates a correct response. Once successful, this specific pathway is saved in memory for future use, demonstrating how the network learns from its mistakes. 

While it's undeniable that artificial intelligence borrows its specificities from the human brain, it also helps biology solve problems it couldn't tackle alone and progress exponentially.

Proteins, the building blocks of life, rely on their unique 3D shapes to function. Determining this structure is a major challenge in biology, traditionally being a slow process. This is where AlphaFold 2, an AI from DeepMind, changes the game. It uses deep neural networks to predict a protein's 3D shape solely relying on its amino acid sequence with stunning accuracy. This capability accelerates drug discovery by pinpointing where molecules can interact with a protein, and it is vital for rapidly developing vaccines and treatments by mapping proteins of pathogens.

An international team, the FlyWire consortium, has successfully mapped the complete neural connectome of an adult fruit fly brain. Led by neuroscientists from Princeton and Cambridge, the project charted over 50 million connections among a staggering 139,000 neurons. This comprehensive map reveals the complexity of the tiny brain, less than 1 mm wide, that enables sophisticated behaviors like learning, flight, and social interaction. Crucially, this milestone was only reached by deploying AI, which single-handedly assembled these fragments into a precise 3D map, revealing a hidden world of connections that was once lost in a vast ocean of data. 

Although artificial intelligence greatly contributes to advances in biology, it is pushing the boundaries between the biological and the digital, raising questions about transhumanism and the limits between human and machine.

The development of a living computer, functioning with lab-grown brains, marks a fascinating and controversial advance that redefines the boundaries between humans and AI. The Swiss company FinalSpark has created a “neuro-platform”, an online computing service where conventional servers based on electronic circuits are replaced by human brain organoids cultivated in a lab. These structures mimic the organization and functions of a human brain, notably its capacity for adaptability; they react to external stimuli and interact with each other to process data or perform calculations. It's no longer humans using the computer, but rather the computer drawing its operation from human tissue.

Ultimately, the relationship between biology and artificial intelligence resembles a two-way street, AI draws from biology its fundamental principles: structure, connections, specificity, even its nomenclature to build itself and evolve. In return, AI opens up unprecedented perspectives, providing biology with tools for analysis, modeling, and innovation. That said, this promising alliance calls for ethical reflection, especially at a time when the boundaries between human and machine are gradually fading.