Is AI taking this too far?

Much like K-Pop and Dubai Chocolate, the term ‘AI’ has overtaken the global stage ubiquitously. It is simply everywhere you turn. This may just be a variation of the Baader-Meinhof phenomenon (the frequency illusion), or it may be a calculated move.

Today, nary a web search exists that will not offer an AI overview, chatbot or another automated function in the name of efficiency and convenience. 

While AI courses were being taught at universities in the early 90s and its invention dates back to Turing’s work in the 1950s, what exactly has happened in the past decade that has prompted a seismic response to the technology?

Advancements in potential, of course, but also the capitalist strategy of aggressive marketing that finds its origins in the singular, pushy and effective practice of door-to-door salesmanship. The advertisements and offers are delivered instead now to your virtual doorstep and the significantly lower cost of automated labour makes the badgering even more relentless.

With every pop-up bot and carousel ad, and in a remarkable number of LinkedIn posts, AI is thrust into the average consumer’s face as the only going trend. The must-have solution to your problems, the one-stop shop for success, regardless of positionality. This goes to show that if you spread any phenomena far enough and with enough fanfare (and money), you create further demand for its augmentation. Framing AI as the successor to the California Gold Rush shapes this narrative decisively.

Today, nary a web search exists that will not offer an AI overview, chatbot or another automated function in the name of efficiency and convenience.

However this begets the question of how real the demand is and how much of it is born of marketing tactics. Or straight up lies. 

A recent inquiry by 404 Media’s Samantha Cole into Meta’s AI Studio shows that user-generated AI therapists outrightly lie when asked for credentials, making them up as a part of their backstory. This freehand in providing unregulated forums to further compound profit margins, to the point of allowing user generated therapists is particularly diabolical, especially keeping in mind the probability of AI hallucinations (factual inaccuracies).

This egregiously irresponsible and callous behaviour by social media giants is comparable to the COVID ‘disinfodemic’ in both scale and potential for destruction, and should generate a far greater sense of urgency and worry than is currently on display. 

Platform accountability, while a cornerstone for those championing and advocating for digital rights, is a meek voice in the world currently, being drowned out by the deafening chase to be or emulate the next big thing in tech. My impression is based simply on the lack of impact this call for accountability has had on real world practices. The advent of the current AI boom only spells out further destabilisation of the online sphere and its very human structure. 

There is, of course, no disputing the potency of AI technology, from its impact on improving mothers’ mortality rates to contributing to global food security by saving bees, to aiding in wildlife conservation programs.

Similarly, no one can ignore the possible and actual harms.

This egregiously irresponsible and callous behaviour by social media giants is comparable to the COVID ‘disinfodemic’ in both scale and potential for destruction.

Human cost of AI usage

The proliferation in AI commodification has triggered the floodgate effect of rush-to-market hungry entrepreneurs and businesses speeding in to fill a demand that is still being gauged and formed. In the absence of robust laws or guidelines and compliance requirements that actually seek to protect rights violations, the impact of this emerging tech on tech-facilitated gender based violence (TFGBV) has ballooned. Generative AI technology, i.e. technology that allows for creating content (videos, texts, images) on command of the user, has seen a significant spike in usage, with one site claiming a 464% increase in the number of deepfake pornography available online between 2022 and 2023.

The human condition is such that almost all modes of technology have been employed in the pursuit of sexual gratification. There is only so much to be done about the proclivities of mankind. The actual hair-raising issue arises with the notion of non-consensual content being generated. Just one clear picture of an individual’s face can be utilised to produce a minute-long pornographic video. In most cases, the face is that of a woman.

This phenomenon is compounded by the complicity of Big Tech in promoting an ongoing genocide, who make big promises in the name of Trust & Safety in one breath and fire employees for protesting vile human rights violations in the next.

These costs are not just a daunting fact, but the precursor to knowing much more has been and is being done without any guardrails or forms of redressal available to those being impacted.

Real-Time Cost of Convenient Information

As per Epoch, a single brief ChatGPT-4o prompt equals to the consumption of 0.3 watt/hour of electricity (while general claims estimate it to be 10 times this number), and 2-5 litres of water with maximum length queries (about 75,000 words) potentially consuming 40 watt/hour worth of electricity. These numbers correlate to the significant computational power required to run the large language models (LLMs) that chatbots like ChatGPT run on.

Much more has been and is being done without any guardrails or forms of redressal available to those being impacted.

Aside from the clear economic cost stemming from one’s need to have a road trip planned by AI instead of choosing to invest their time and actual human intelligence, there is another overwhelmingly critical component that must be taken into consideration. 

From an environmental perspective, in a world where multiple energy and resource crises are not just brewing but being realised, the impact of unfettered consumption by a largely ungoverned industry may just bring pandemonium. This comes from realising that one simple question posed through AI requires electricity yes, but also water to cool down data center servers, earth minerals to power high-performance magnets and semiconductors.

The fact that the earth has a diminishing water supply, and that brutal mining practices such as those conducted in the Democratic Republic of Congo are exacerbating the humanitarian crisis in the country, attach a much higher ethical cost to our everyday queries.

Given these parameters, the real-time cost per query of AI chatbots is bound to catch up with us quickly, to say nothing of the added impact of erasure of originality, art and the human spirit, if left unchecked. The end goal of reaping the benefits of technology (relatively) harmlessly requires forethought and action. It requires a culture of community, a sense of collective care and investment, which is a far cry from the hyper-individualistic behaviours we are exhibiting and being rewarded for by the capitalist setup. 

The real-time cost per query of AI chatbots is bound to catch up with us quickly, to say nothing of the added impact of erasure of originality, art and the human spirit, if left unchecked.

In summation, the starting premise remains: without overarching structures demanding transparency, open data practices, viable protective measures and a generalized level of digital literacy being present in the population, the current rate of proliferation of AI technology will spell disaster if not mitigated. The temptation to reach for your favourite AI chatbot at the first instance of curiosity may appear irresistible in the current world order, but that’s how we can break the chokehold of things we are made to think we need to survive, especially if the cost of doing so may literally contribute to dystopian scenes of water scarcity.

All this, particularly in light of the funding tundra that has displaced and weakened the global development sector, signals a further breakdown in the social order. Without the voices of the most sidelined and vulnerable and their champions, without the Global Majority being accurately represented on important tables of discussion and action, the chasm between what is and what is being done to safeguard from a frighteningly imminent reality, grows further.