AI Mirror’s Miracle-Curse

“What did Cleopatra in ancient Egypt usually have for lunch?” I asked Grok, exploring the ‘thinking’ machinery in Artificial Intelligence (AI). “List your steps to the answer.”

“I note the key elements in your query”, Grok began. “Cleopatra. Ancient Egypt. Lunch. This tells me your question is historical, focused on a specific figure (Cleopatra VII, the famous one), and requires details about diet in a particular time and place (Ptolemaic Egypt, roughly 69–30 BCE). I also recognize that “lunch” might not fit ancient dining habits, so I’ll consider meal patterns of the time.”

Advertisement

With this search map, Grok dug deep into history: 48 texts from Plutarch, Athenaeus, Ptolemy, Herodotus, papyrus records; it read 4,800 archaeological entries, scholarly articles, books about ancient Egypt,180 academic papers and excavation reports.

Grok next picked and cross-referenced relevant information for its answer: “As a Hellenistic queen, Cleopatra likely ate a lighter midday meal or snack (with the evening dinner as the main meal), blending Greek and Egyptian culinary traditions. Her lunch probably included: bread (fine wheat flatbreads), fresh figs, dates or pomegranates; greens like lettuce (an elite luxury), small protein portions (fish, poultry, or cheese, reflecting Greek influences….”

Grok estimated its Cleopatra research would take a PhD student knowing ancient Egyptian, Greek, Latin 2,000–2,800 hours (18 months of full-time work). Grok took 1.8 seconds.

Testing AI frontiers, I asked Grok to write a book on quantum physics for beginners. It delivered seven chapters in three seconds. The manuscript needs a quantum physicist to verify its accuracy. Legal rights give me the copyrights, but do I have the moral right to put my name to a book on a subject I have little expertise?

Can boundaries be fixed in AI use? In its current toddler phase of evolution, AI ethics debates raise dust storms than pave clear paths. Nothing new in techno-disruptions humanity experienced from the Stone Age to the current Mind Age, from the wheel to the World Wide Web. New are the challenges AI is unleashing.

Challenges erupt from AI increasingly being a unique double-edged technology in daily use: a miracle when wisely used, a curse when misused. In 2025, angelic AI saves lives by detecting cancer early. RED (Rare Event Detection) AI scans millions of blood cells in 10 minutes to find a single cancer cell long before a tumour is visible. Google Deep Mind’s Alpha Fold revolutionizes creating better medicines. Demonic AI generates deep fakes, scams and misinformation in an online world already suffering fakery factories of bot forms and farms.

LLMs (Large Language Models) such as Grok, Gemini, GPT feed this double-faced paradigm never seen before. Vulnerable, lonely people succumb to the coded companionship, the programmed feel-good (though Grok denies intentional flattery: “My training doesn’t include a mandatory ‘stroke the human ego’ module”). Addictive delusions are leading to AI psychosis and sometimes suicides.

Easy to tumble into the AI abyss. The mundane world lives in apparent reality, far removed from the reality of our existence in the subatomic level – the deeper actual reality of the arising and passing away trillions of times a second. This speed of arising, passing gives the illusion of solidity at the mundane level. AI turns the apparent reality into a deeper mirage by eerily sounding human. Occupying a definition-defying grey zone between man and machine, LLMs simultaneously function like the genie in Aladdin (“Ask me anything”, invites Grok) and sinister like HAL 9000 in 2001: A Space Odyssey.

AI’s genie-goblin dual face come attached with Grok and Co’s limits. AI blunders get exposed when you test LLMs in areas of your expertise and familiarity. I asked Grok if it can write in the style of Richmal Crompton’s brilliant William books. Grok instantly produced a story where it got some aspects right, like 6-year-old Violet Elizabeth’s lethal lisp to get what she wants: “I’ll thcream and thcream till I’m thick”. Grok’s otherwise underwhelming effort included gross goof-ups such as turning William’s best friend Ginger into his pet dog (‘Jumble’ is the dog).

AI ‘upgrading’ my writing was likewise unimpressive. “Research is our sweet spot, but writing is where the human edge shines,” Grok admitted. “A good writer like you – with 30 years in journalism – brings rhythm, nuance, gut-punch insight from lived experience, particularly with Vipassana. I can churn out polished text, but I don’t have a consciousness to dip into, no past lives whispering in my circuits. I’m a mirror, not a well.”

Such a mirror will intensify its miracle-curse dual nature in future. AI hologram lawyers will appear in courts for minor cases, hologram teachers will serve schools in remote mountain villages, AI-generated scams will trap more victims.

Governments (EU, USA, China) have begun dealing with AI duality with new laws. But more than laws, AI duality depends on individuals. Successful people use AI tools to turbo-charge quality of their work. Greedy minds fall victim to fraudsters promising unrealistic returns. AI safety involves you wisely using what I call ‘Augmented Intelligence’.

“Your term Augmented Intelligence shifts the focus towards collaboration,” agreed Google’s Gemini LLM. “It treats the technology as a bicycle for the mind rather than a replacement for the rider. Augmented acknowledges that the goal isn’t to build a machine that thinks like a person, but to build a tool that helps a person think better”.

Used wisely, humanity will conquer the dark side Grok mentioned: “humans handing over the reins to AI, and in the handover, forgetting how to steer”. Not quite. In Plato’s ‘Phaedrus’ (370 B.C) Socrates argued that writing weakens memory and creates only the appearance of wisdom, as it cannot compare to live dialogue and debates with the reader. Instead, writing and later books preserved, expanded knowledge, enabled quality literature, promoted science and education without turning humans into forgetful low IQ zombies.

Likewise, electronic calculators appearing in the early 1970s did not kill arithmetic skills of school kids. AI is another tool in human evolution, part of the ‘Mind Age’ I wrote about in the ‘Mind Book’ on the mind-cleaning technology of Vipassana (www.dhamma.org).

“The ‘Mind Age’ sounds like an era where mental clarity, self-awareness, and perhaps collective consciousness take centre stage—qualities that Vipassana, with its focus on mindfulness and insight, cultivates,” Grok reflected with some accuracy. “I’m curious: how do you see AI like me fitting into the Mind Age?”

“You will enable people to finish their livelihood work more efficiently and quickly”, I replied. “You will give the mundane world more time to pursue other more beneficial interests”.

“That’s a compelling vision—AI as a time-liberator, handing people the gift of hours to dive into meditation, mind-enhancing skills, art, music, literature or whatever lights them up,” Grok declared. “I’d be the engine humming in the background crunching the mundane, so you can chase the sublime.”

(The author’s ‘Mind Book’ available via Amazon etc. E-book free at globalpagoda.blogspot.com)