Passé

On AI, existential risk, and the species that never learns

"Science should not know when to stop, engineering must." Sambhav Siddhartha Varshney

Now that we have started the blog with vainglory, why not continue? If you believe I have set a precedent for pomposity, another blog will not hurt. I accept that a lot of you (if anyone still reads these blogs) think of these as a medium for me to vent out all my cynicism and bitter view of the world. Believe me, this time while writing the blog, I do too.

My limited understanding of psychology from the three books on the subject I have read so far has imparted to me enough wisdom and strength to acknowledge my shortcomings, biases or flaws publically; the obvious bias here might be against artificial intelligence (AI), leaving us (the peasantry, less computer science savvy) economically and/or socially redundant in the long term, masquerading as my irrational fear of AI becoming the cause for our extinction.

Until recently, for me, the primary cause of concern that would lead to the extinction of all species on the planet was that of climate change. Luckily for it, there is a new competing scapegoat in town for the 400 metres sprint, the AI.

Now make no mistakes, both of them are just as willing to run the race as do horses in their analogous counterpart. Until the occasional horse gets a kick out of it.

I would rather let my species get extinct eating and enjoying cheese hamburgers, butter chicken, milk and downloading vast amounts of Netflix content than waiting for OpenAI to increase the server capacity for all of us to roam simultaneously like little toddlers with their RC cars, around the fields of ChatGPT looking for some rather racist, stupid (relatively) fun which might just casually render huge swathes of hardworking and talented humans, unemployed.

It has never been my intention or philosophy of life to censor or squash speech, ideas, and human endeavors regardless of their associated terrains. Writing this blog, I almost feel at odds with myself and related lifelong ethos. I firmly believe any discovery or invention that provides extensive instrumental services to their human lords is worthy of use, regardless if there are a few annoying obstacles in the way. The internal contradiction here is the point. If you're not at odds with yourself, you're not thinking hard enough.

The eminent physicist and professional tongue model* *Pun shamelessly borrowed from Diane Morgan's Philomena: Cunk On Earth TV Show. Albert Einstein jotted down, perhaps the most famous scientific equation of all time — $E = mc^2$ — from which we reap immense benefits to this day. I sincerely believe, as time passes and the theories of relativity, especially that of general, make more sense and coalesce with other major physical descriptions of reality, the profound realizations from the two theories of relativity, will grow manifold.

But they did not always meet the path to this enormous acclaim with acceptance. Einstein admitted to heartfelt regret and grief post incidents of nuclear bombing in Hiroshima; had he not brained his way through the space-time fabric of reality, hundreds of thousands would have not perished in thermal hell.

"That dwarf Kim Jong-un finna been dead had not been for my nerd ass." Albert Einstein. (This quote is fabricated. Obviously.)

Apologists for free-flow of AI often cite nuclear energy and its growing role in solving anthropogenic carbon crisis and draw parallels between the two revolutionary scientific/engineering stalwarts and like to categorise the nuclear plant blunders, bomb threats and potential AI dangers and tragedies as preliminary necessities; inevitable trade-oops as part of growth which will flatten out and stabilize as our understanding of the subjects stockpiles.

Although this sounds like a reasonable argument, I insist it is not and for several valid reasons.

The first and the most obvious, in the face would be the sheer scalability of AI — unlike nuclear weapons of mass destruction (WMD), AI can be scaled to literally every street on the planet through its ever-growing fleet of engineers and researchers writing millions and billions of lines of code each passing day. Them training their models on quadrillions of parameters, it would be as difficult to control and regulate every one of them as it is to stop cyber criminals and illegal activities on the dark-web despite trillions of dollars invested by law enforcement agencies and internet security companies, which implies there would be a regulatory body, which I doubt. The deep pockets and influence of big-tech hold enormous power and influence to keep lawmakers at bay.

They allow and create a smoke screen environment of 'dissent' that encourages face-value shunning of their growing Orwellian policies to keep shallow, less knowing dissenters at bay. The arduous ones that stay are in such low numbers, to the big-tech it is hardly bothering.

I understand all this sounds rather like the sort of blabbering that comes from chatter pieces of conspiracy theorists, but the naïve inside me has subsided a little. The philosophers of science I so fondly liked to diss and look down upon in the past have historically made little sense to me. The arrogance and the monochromatic world view that scientific efficiency is ubiquitously good has cast disastrous consequences for us, and the impending doom that stems for that of artificial general intelligence might just be the last we ever have to suffer. Eternal peace in the end, though, a little reassuring?

Another valid reason would be stigmatisation. Even though there have only been two major nuclear power plant disasters, and their impacts have turned out to be less severe than initially expected, the stigma around these behemoths remains stable. Despite growing evidence and promotional campaigns by nuclear power activists regarding the safety and benefits of the power source, the optics of a widespread destruction cease to go away. With AI, although some stigma is surely present — thanks to movies like Terminator and early stage advocacy for AI regulations by established and renowned innovators, experts, public speakers and thinkers on the topic like Stephen Hawking and Elon Musk — the rational concern is nowhere near justified and required levels.

The instant sense of awe and joy generated by an AI powered chatbot or image generator usurps that derived from the discovery of most other inventions, particularly in the academia, for they need extensive experience and strong academic experience in their respective fields and applications are not exactly well-defined or achievable because of several constraints, such as ease of use and economic feasibility.

But the booming tech industry knows of no such hurdles, draped in several trillion-dollar market capitalised international conglomerates twined with emphatic reach and demand, bringing home returns on these investments is rather easy. Just see what Microsoft's $10 billion dollar investment in OpenAI has brought to the firm. Having now integrated their grossly under used and lackluster search engine Bing, the demand for it has suddenly exploded. The perennial search engine behemoth Google broke the sweat for the first time in decades! But now they are running the race too.

The prestige and money that comes for AI engineers today has already become the 80s equivalent of MBBS and B.Tech degrees. It was as if one could hold themselves in the same regards as kings of the 17th century India. And it will only boom, given the high-pay the profession guarantees, candidates from all fields of engineering have already semi-migrated to the field and it is only about time those from science, medicine and humanities would ultimately have to as well.

Once a sizeable chunk of professionals and human machinery drifts towards automation and related jobs, it would be a disaster to suddenly decommission and layoff the tech and the employed if things turn out to be a little M3GANy. I highly recommend you to give the movie a watch.

"I'm the primary user now." M3GAN, Antagonist — Film: M3GAN (2022)

Occurrences and reportings on the merged Bing-ChatGPT platform getting contentious have surfaced. Although I do not know enough about the technicalities of what caused it, whether it is an anomaly or just an insight into what expectations to set for the future. It might be a sneaky attempt by Microsoft to gather traction for all it matters. Rather safe than sorry.

But we will reserve for ourselves the benefit of the doubt and take cues from these instances. Despite, and it a very burdensome 'despite' for me to type, almost cynical, the promise of all that AI offers — and again I do not just type 'the promise of all that AI offers' as a snide, ungrateful remark. I fully acknowledge the potential of the technology.

Current champions of the univalent altruist nature of AI rely irrationally and heavily on human capability to avoid the urge to do the wrong and the stupid. To resist the wrong might come to most of us, but the valor to fight the stupid? One look at any history textbook would tell you this is an audacious observation to make. So, given our track record of stupidity and greed, my bet would be against the utilization of AI in more beneficial than harmful outlets. Even if it turns out to be wrong, quantity might not always have the last laugh. One high-quality f*ck up could cast the end spell for us all.

"Two things are infinite: the universe and human stupidity; and I'm not sure about the universe." Albert Einstein

The creation of robots that can do for us the tasks that we deem puny or lethal. Machines that can reduce human errors to zero. Aid physicians and surgeons in diagnosis and treatment of complex ailments. Run simulations to learn methods to oust cancers from society just like polio and so much more. I would have to turn to Michio Kaku here to enlist them all, but I find myself to be more in line with Christopher Hitchens for the occasion, so make do with all the pessimism you can catch.

It always came to me as cringe whenever art romantics made disparaging comments on the scientific quest for the absolute: telling us that perfection draws the joy out of human experiences and that human life would become dull without subjective experiences. Slowly, but surely, this has marked a profound change of mantle in my world-view.

"The genie is out of the bottle. We need to move forward on artificial intelligence development, but we also need to be mindful of its actual dangers. I fear AI may replace humans altogether. If people design computer viruses, someone will design AI that replicates itself. This will be a new form of life that will outperform humans.

Success in creating effective AI could be the biggest event in the history of our civilization or the worst. Perhaps we should all stop for a moment and focus not only on making our AI better and more successful, but also on the benefit of humanity." Stephen Hawking