In February 2020, a team of researchers at MIT gathered around a set of screens showing the output of a deep-learning model. They had asked the system to search a vast chemical universe, millions of compounds beyond what any human chemist could feasibly consider. Out of that ocean, the algorithm produced something startling: a molecular structure with no known precedent in pharmacology, yet with the fingerprints of a powerful antibiotic. When tested in the lab, it killed drug-resistant bacteria that had defied decades of human research. They named it Halicin.
The mood in the room mixed awe with unease. Here was a drug candidate discovered not by a scientist’s hunch or painstaking trial-and-error, but by a machine running through possibilities at inhuman speed. For some, this was proof of a new frontier. For others, it raised an unsettling question: had the machine invented something? Or had it simply performed a glorified search, the digital equivalent of throwing dice until chance produced a winning throw?
That question has haunted computing since its inception. Nearly two centuries earlier, Ada Lovelace, looking at Charles Babbage’s designs for his Analytical Engine, wrote that the machine “has no pretensions whatever to originate anything.” Lovelace saw machinery as bound by human input, tools that could amplify calculation but never imagination. The true work of invention, she believed, remained in the human mind. Alan Turing, writing in 1950, disagreed. He suggested that machines could surprise us, and in those surprises lay the seeds of originality. To him, the distinction between calculation and invention was less rigid; novelty itself could emerge from the machine’s processes, even if the machine had no awareness of what it had produced.
The tension between Ada’s skepticism and Alan’s optimism has only sharpened as artificial intelligence systems move from toys and tools to collaborators and rivals. The NASA antenna that looked like a twisted paperclip, designed by evolutionary algorithms and launched into orbit, carried the debate into space. The discovery of Halicin carried it into medicine. DeepMind’s AlphaFold, which predicted more than 200 million protein structures, carried it into the very architecture of life. Each breakthrough forces the question anew: are these inventions, or only by-products of human programming?
Neuroscience complicates the picture. Invention in the human brain is not a simple act of creation ex nihilo. It arises from recombination, from the collisions of memories and associations under the influence of curiosity and context. The “eureka” moment is less divine spark than neural coincidence, a cascade of patterns suddenly aligning. In that sense, what machines do—recombining massive amounts of data, searching for patterns that yield novelty—looks uncannily similar to our own process, albeit without consciousness or desire. Critics point out that machines lack understanding, that they cannot assign meaning to what they generate. Yet even in human history, many inventions emerged from accidents or blind processes: penicillin, discovered by chance mold, or vulcanized rubber, stumbled upon in a messy lab experiment. In those cases, the human role was not origination but recognition, the reframing of an accident into an invention. If we grant ourselves that leeway, should machines be denied it?
The unease is not merely academic. In 2019, Stephen Thaler’s AI system, DABUS, generated designs for a food container and an emergency beacon. Thaler filed patent applications naming the system itself as the inventor. Patent offices in the United States, the UK, and Europe rejected the filings, insisting only humans could be inventors. South Africa granted one; Australia’s courts briefly sided with Thaler before reversing course. The spectacle revealed how unprepared the law is for this new frontier. If machines are inventors, who owns their creations—the programmer, the operator, the corporation, or the machine itself? If they are not, are we discouraging innovation by refusing to recognize genuinely novel outputs?
Culture, too, is wrestling with the implications. In art and music, AI-generated works draw both fascination and suspicion. A portrait painted by algorithm sells at Christie’s for hundreds of thousands of dollars, while neural networks churn out symphonies that echo human composers. Some dismiss these as parlor tricks, others as theft. But the discomfort echoes past upheavals. The printing press threatened memory, photography threatened painting, industrial looms threatened artisans. Each time, human creativity adapted, reshaping itself around the machine. The Alan–Ada debate may be less about the death of invention than its redefinition.
Reddit threads and online forums give this debate a raw, unfiltered voice. Some argue, echoing Lovelace, that AI cannot create new knowledge, only remix what already exists. Others, channeling Turing, point to discoveries like Halicin or AlphaFold as proof that the machine is already operating beyond human foresight. A few speculate that true invention may only come once machines are embodied, equipped with sensors and actuators that let them experiment in the world as humans do. If invention is tied to interaction, to the feedback of trial and error in the physical environment, then the first truly inventive machines may be robotic, not purely digital.
What unnerves many is the alien quality of machine outputs. AI-generated designs often look organic, even grotesque, outside the aesthetic comfort zone of human intuition. And yet they work, sometimes better than anything we could produce. Invention has always been defined not only by novelty but by usefulness. If usefulness can be achieved without comprehension, then perhaps the bar for invention has already been crossed.
This tension runs deep into our conception of what makes us human. For centuries, invention has been one of the pillars of our uniqueness, a source of cultural pride and identity. To share that role with machines feels destabilizing, a challenge to our sense of agency. But perhaps the future is not about surrender but about partnership. Machines may generate the alien possibilities; humans will assign meaning, ethics, and context. The act of invention may become a dialogue rather than a monologue, an expanded ecosystem of creativity where surprise comes from silicon as well as flesh.
The Alan–Ada debate endures because it asks not only whether machines can invent, but also what we mean by invention itself. If we define it as intentional creation with understanding, Lovelace’s caution still holds. If we define it as the generation of novelty that proves useful, then Turing’s vision is already here. Perhaps the most radical invention of this era will be neither a molecule nor a machine, but a new conception of invention itself—one that stretches to include partners we never expected, and one that forces us to rethink what it means to originate anything at all.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
The Alan–Ada Debate: Can Machines Invent?
