According to SHARE Kansas City session #10526, “What Every z/OS System Programmer needs to know about AI” with Steve Warren and Meral Temel (both of IBM), “AI refers to the ability of computer systems to attempt to mimic the problem-solving and decision-making capabilities of the human mind.”

This definition has evolved slightly since 1950, when Alan Turing proposed the Turing test for artificial intelligence, originally called the imitation gameⁱ. He described AI as a machine's ability to exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human. However, he didn’t yet call it “AI,” referring instead to the question, “Can machines think?” Six years later, the field officially known as artificial intelligence came into being^ii,iii.

But the question of whether machines created by humans could exhibit thinking behavior goes much further back in history, even if its current name is relatively recent.

Turing proposed that the only measure of whether a machine was thinking was whether it interacted in a manner indistinguishable from that of a reasonably intelligent human. This is philosophically very different from the question of whether a machine is actually thinking, which we seem to have mostly set aside for now in pursuit of a good simulation — or perhaps more accurately dissimulation.

But if we rewind through history, we quickly discover that we have wanted to design human-made creations that behaved as if they were thinking like humans for a very long time. So, rather than document advances in the technology since the 1950s, perhaps it’s time to take a chronological zoom back into the past to dig into the nature of our desire to create inventions that behave like they are thinking, ponder our motivations, and look at the outcomes in the light of the modern utilitarian focus on results rather than more esoteric considerations such as provable consciousness.

As we gaze at the tapestry of human engagement with artificial thinking-like behavior, there are three important threads that emerge as weaving along the axis of time: fiction (including science fiction), theoretical intentions, and practical inventions.

One of the theoretical intentions has been to create a mechanism that thinks like a human without having to engage in the traditional biological process for procreating new humans, whether working with the gears and wires of industrial era technology, or building with previously existing biological materials, or employing more mystical means (keeping in mind the famous quotation that “Any sufficiently advanced technology is indistinguishable from magic” — one of science fiction author Arthur C. Clarke’s three laws^iv). In the world of fiction, Frankenstein’s Monster as written about by Mary Shelley in 1818^v, is an early modern prototype of such a construction. But in terms of industrial era machines that exhibit thought-like behavior, one might look at Jacquard’s loom, the first punch-card-programmable machine^vi. Parallel to this genuine innovation was the famous Mechanical Turk, also known as the Automaton Chess Player^vii, which mystified audiences with its apparent ability to play chess for 84 years until exposed as a fraud by the son of the machine’s owner and by Edgar Allan Poe^viii.

Of course, this calls to mind the important question about all artificial intelligence: to what extent is it a sophisticated puppet, and to what extent is it truly able to be self-determining — reminiscent of the Wizard of Oz^ix, whose puppet master Professor Marvel famously said, “Pay no attention to the man behind the curtain!” Funnily enough, we may well ask that question, but then discover that in some ways it is a more political than technical question, as humans with ostensible free agency are known to behave in a manner that manifests the intentions of others as well. Likewise, a machine that behaved in an acceptably intelligent manner would have to be in some way a manifestation of the wishes of those who created and/or funded it.

This may remind us of another conundrum from the world of theoretical AI: the Chinese Room thought experiment^x, in which a nearly-infinite dictionary of sentences and responses has been contrived for a machine or human who doesn’t speak Chinese to look up any question or comment and respond accordingly per the dictionary. Fortunately, whether this is indeed as preposterous as I believe it to be or not, the only insights it offers on Turing’s test is the idea that there is more than one way that it might be gamed, at least given a sufficiently gullible examiner.

Panning through history, it is instructive to note that, in addition to industrial-era mechanisms being envisioned as AI-like creations, there were (apparently fictional) stories of human-made biological entities such as homunculi^xi, artificial human-like beings created through alchemy. And there were stories of human-made statues, idols, and puppets which were brought to some semblance of sentience. Three examples of these are: golems, which were ostensibly animated by scripts or incantations^xii; Pygmalion, a myth about a statue being given life by a divinity^xiii; and Pinocchio, a wooden puppet who was first animated and then brought to life by a fairy^xiv. Parallel to and sometimes even implicit in these stories is the ever-present question of fraud, such as the idol that apparently ate food offered to it in the deuterocanonical story of Daniel and Bel^xv.

As we dig deeper and deeper into history, it begins to look like our motivations for creating artificial intelligence include creating something in our own image — a manifestation of the idea that humans were also created in the image of a creator.

Yet our motivations have been much more variegated over the eons, and if these included creating an easy-to-manage workforce or machines that could think bigger thoughts than their creators, there have also been concerning elements of avoidance of responsibility and making illicit gains through fraud.

Returning to our modern era, these illustrations of our eternal journey of trying to make creations in the image of our own intellects remind us that, while there will always be new innovations and insights in the world of AI, and even new theories of how to create, test, and validate them, the one thing it all has in common is that it is a deep expression of our humanity. And so, as we look for ways to evaluate the validity and relevance of such advances, the essential question of whether it truly affirms our humanity is perhaps the most important thing to take away from this article.

Find more education and webinars about AI and other topics in SHARE’s webinar library, accessible to members of SHARE. Learn about becoming a member and the benefits you’ll receive today.

Sources:

i. https://en.wikipedia.org/wiki/Turing_test

ii. https://en.wikipedia.org/wiki/Computing_Machinery_and_Intelligence

iii. https://en.wikipedia.org/wiki/Artificial_intelligence

iv. https://en.wikipedia.org/wiki/Clarke%27s_three_laws

v. https://en.wikipedia.org/wiki/Frankenstein%27s_monster

vi. 1804, https://en.wikipedia.org/wiki/Jacquard_machine

vii. https://en.wikipedia.org/wiki/Mechanical_Turk

viii. https://en.wikipedia.org/wiki/Maelzel%27s_Chess_Player

ix. https://en.wikipedia.org/wiki/The_Wizard_of_Oz

x. https://en.wikipedia.org/wiki/Chinese_room

xi. https://en.wikipedia.org/wiki/Homunculus

xii. https://en.wikipedia.org/wiki/Golem

xiii. https://en.wikipedia.org/wiki/Pygmalion_(mythology)

xiv. https://en.wikipedia.org/wiki/Pinocchio

xv. https://en.wikipedia.org/wiki/Bel_and_the_Dragon