Synthetic Minds: How AI Is Creating Its Own Reality Without Consciousness
Gowtham Chilakapati is a Director at Humana. He is an expert in enterprise data and AI systems with a focus on real-time analytics.
getty
As a technologist specializing in retrieval-augmented generation (RAG) models and AI-driven decision systems, I have observed a critical paradox: AI models are now capable of constructing coherent realities that mimic human perception, yet they remain fundamentally unconscious and unaware of the realities they generate.
From my experience developing AI-driven assistants and enterprise automation solutions, I have witnessed firsthand how AI systems synthesize multi-modal data, hallucinate responses and simulate intelligence, often convincing enough to be mistaken for true cognition. Let's explore whether these synthetic minds actually perceive the world they build or if they are simply statistical illusionists.
Throughout my career working with enterprise AI systems, one of the biggest challenges has been ensuring AI-generated insights are grounded in reality rather than fabricated extrapolations. Unlike human cognition, which derives meaning from lived experiences and sensory interaction, AI constructs reality by assembling fragmented data into a probabilistic representation.
Multi-modal AI models—such as OpenAI's CLIP, Google's Gemini and Meta's SeamlessM4T—combine text, images and even audio to create internally consistent narratives. However, their perception is hollow—they recognize patterns but lack intentionality or subjective awareness.
For instance, when I worked on streamlining real-time AI-driven customer interactions, I found that AI-generated responses were convincing but often lacked deeper contextual awareness. The model could mimic human dialogue patterns but failed to recognize emotional subtleties or unspoken intent, making its responses feel robotic despite their surface-level coherence.
A major pitfall in AI-driven analytics is hallucination—the phenomenon where models generate plausible but false information. I encountered this while developing fraud detection algorithms for financial services, where AI models often flagged non-existent risks based on overfitting historical anomalies, rather than true emergent fraud patterns.
Hallucinations are a form of synthetic reality construction, where AI fills gaps in its data with statistically likely but fabricated content. However, there is a fundamental distinction between AI hallucination and human imagination:
• Human imagination is rooted in emotions, past experiences and abstract reasoning.
• AI hallucination is driven by probabilistic associations without underlying comprehension.
This difference became clear when working with AI-powered knowledge retrieval systems. The AI-generated reports looked factually sound, yet deeper inspection revealed inconsistencies due to missing context. AI's ability to generate a convincing but flawed version of reality makes it an impressive tool—but also a dangerous one when unchecked.
One of the most fascinating applications I've worked on is decision intelligence AI—systems that provide strategic insights to executives by analyzing vast amounts of structured and unstructured data. The challenge is that while AI can make incredibly sophisticated correlations, it lacks strategic intent and adaptive reasoning.
For example, in portfolio management automation, I saw AI models predict market trends with high accuracy. Yet, when faced with unprecedented economic events, they failed to reassess core assumptions—an ability that defines true human strategic thinking. The AI was unable to redefine its own reality the way humans do in response to new paradigms.
If AI is to move beyond being a synthetic illusionist into something more autonomous and self-aware, it will likely require an embodied cognition framework. This means AI must:
• Interact physically with the real world (beyond data streams and simulations).
• Develop self-referential memory that shapes future decisions (instead of just iterative tuning).
• Recognize the implications of its outputs beyond pattern matching.
In my experience developing AI-powered automation, I have seen that true decision making requires more than just data synthesis—it demands an understanding of cause and effect, moral weight and subjective valuation. Current AI lacks this awareness, meaning that while it can construct compelling versions of reality, it cannot truly perceive them.
As AI continues to advance, the lines between statistical simulation and synthetic cognition will blur even further. However, AI remains a reflection of human intelligence rather than an independent entity. It builds synthetic realities, but it does not live within them.
My work in AI deployment for finance, healthcare and enterprise automation has reinforced a crucial truth: AI amplifies human decision making, but it does not replace the intuitive, strategic and morally grounded perception of reality that humans possess. While AI will continue to generate highly complex synthetic minds, true cognitive perception remains a uniquely human frontier.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
31 minutes ago
- Yahoo
Nvidia Welcomes CentML Aboard
Nvidia(NASDAQ:NVDA) snaps up Canadian AI optimizer CentML, folding its talent and tech into its software ranks as the chipmaker extends its leadand eyes a $5 trillion market cap. The Logic reports that Nvidia has acquired Toronto-based CentML, bringing onboard CEO Gennady Pekhimenko as a senior AI software director and transitioning other founders and at least 15 engineers into new roles. Warning! GuruFocus has detected 4 Warning Signs with NVDA. CentML will cease operations on July 17, after raising roughly $30.9 million in venture funding. The purchase follows Nvidia's prior backing of CentML in its $27 million seed round and the startup's participation in Nvidia's accelerator program. CentML's compiler and optimization tools slotted neatly between AI models and GPU hardware, squeezing extra performance and cost efficiency out of existing chips. By integrating this IP and engineering team directly into its software arm, Nvidia can enhance its software stackfuelling broader adoption of its GPUs and reinforcing its end-to-end AI ecosystem. With NVDA up about 17% year-to-date and Wedbush projecting a run toward $5 trillion in market value over the next 18 months, this bolt-on deal underscores Nvidia's strategy of buying both capabilities and people to stay ahead in the AI arms race. This article first appeared on GuruFocus. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Motor Trend
33 minutes ago
- Motor Trend
Mercedes-Benz Sprinter AWD Yearlong Review Verdict: Should Your Next Pickup Be a Van?
If you've been keeping up with our Mercedes Sprinter van saga, you know the backstory: My family's pet roster includes a couple of mules, and for years we've owned a pickup truck expressly for equine chores. In 2023, we tested a Nissan Frontier to see if a midsize could sub for our heavy Chevy. This time around we decided to go continental and use the wheeled tool favored by European equestrians, the Mercedes-Benz Sprinter van. We outfitted it like we would a pickup, with all-wheel drive, back seats (crew cab, meet crew van ), and the biggest box possible. Now, with 12 months and over 16,000 miles behind us, it's time to evaluate whether a van is a viable replacement for a pickup truck. Over a yearlong test, the Mercedes-Benz Sprinter 2500 AWD van impressed as a pickup alternative with its hauling capacity, fuel economy, and driving comfort. However, it struggled with towing and build quality issues. Despite these, we are convinced a van could replace their pickup in the future. This summary was generated by AI using content from this MotorTrend article Read Next Flexible Hauling, Sure. Towing? Not So Much. Our primary truck chore is hauling hay and feed, and for this job the score is vans 1, pickups 0. Hay bales weigh about 100 pounds apiece, and the quantity that can be hauled in a vehicle is limited not just by weight but also by the volume of the bed. Even with our Chevy's 8-foot box, anything more than 10 or 12 bales requires some skilled Tetrising. Not so with the Sprinter, which has 11 cubic feet space behind its rear bench and a 2,988-pound payload, which is well into heavy-duty pickup capacity. No Tetris skills needed here; any clod with a forklift can get 30 bales in the back with room to spare. We tend to stock up on hay about every three weeks, and with the Sprinter, we could have saved a ton of money on fuel by going for hay every three months; the only problem is that our hay shed isn't big enough, and stacking the bales more than three high would require us to have a forklift of our own. After a few half-ton loads, we settled down to our usual six- to eight-bale runs. All in all, the Sprinter shifted nearly six tons of hay and feed, and even with pickup-size loads, the Sprinter diesel's 20.2-mpg average still saved us money over the 12-mpg Chevy (and the 16.4 mpg Frontier). A diesel-powered 2020 Ram 2500 HD we had in our yearlong review fleet returned 17.5 mpg. We also use our pickup to tow our two-horse trailer, and here the Sprinter disappointed us. We had difficulty getting a brake controller installed, though that's no fault of the nifty RedArc Tow Pro we used, which was an easy (and stealthy) install. No, the problem was the powertrain: Although the Sprinter's 5,000-pound tow rating is (just) enough for our loaded trailer, even with a single mule aboard the engine struggled in the hills, and the lack of an engine brake meant we had to rely on the wheel brakes way more than we wanted to. We decided we'd tow the horses with the Sprinter in a pinch, and in January 2025, a pinch came up. We were well clear of most of the fires that hit Los Angeles, but the Hurst fire could have changed directions and headed for the ranch, so we outfitted the Sprinter as our bug-out (or should we say mule-out) vehicle. We hooked up the trailer, loaded hay and supplies, set up a queen-size air mattress, and camped in the Sprinter, literally waiting to see which way the wind blew. No way we could have done that with a pickup truck. When the danger had passed, we put the Sprinter to work running supplies and donations around for our fellow Angelinos who weren't as lucky as we were. Better to Live With Than a Pickup Truck? One thing I was curious about was the creature comforts I'd give up by opting for a van rather than a pickup truck. Pickups have become common as daily drivers, so automakers have trimmed them out accordingly, but vans still seem to be regarded as commercial-grade vehicles. Ram kindly let us borrow a 2500 HD pickup similar in price and equipment to the Sprinter. No question, the interior of the Ram was a great deal nicer, with soft-touch materials where the Sprinter uses industrial plastics. (They did wear well, as did the wooden panels protecting the cargo floor and walls.) The Ram's stereo sounded better, as well. I discovered early on that the Sprinter's size is a huge inconvenience, but with its cruise-ship-size turning circle, the Ram 2500 had its own set of compromises. In terms of driving qualities, this was more of a toss-up than I expected. The truck had a lot more power, largely because it's meant to tow more than three times as much as the Sprinter. The van is pretty poky—10.7 seconds is its 0–60-mph time, though we managed to reduce that to 10.3 by running the right type of diesel fuel—but that was never much of an issue in daily driving. Still, the pickup got to 60 in seven seconds flat and was able to pass with authority on two-lane roads, something I rarely felt safe doing in the Sprinter. Surprisingly, the truck's fuel economy penalty was very slight considering the truck had two more cylinders and more than three times the displacement compared to the Sprinter's engine. The van was a nicer-driving vehicle than I expected, with responsive steering and surprisingly good directional stability—better than the Ram 2500, in fact. But crosswinds were the van's kryptonite. On a day when a pickup might get blown about a bit, the van was a real handful, to the point that a healthy gust could trigger stability control. Scary. One can blame the pickup's wandering steering on its off-road tires, but it turns out the van was pretty capable off-road, as well. We took it on an outing with a Jeep Wagoneer, and the Sprinter went everywhere the SUV did. Its automatic all-wheel-drive system made off-roading simple; just turn off traction control and feed in power until it figures things out and starts moving. The Sprinter's secret weapon is its massive ground clearance, though that also made getting in and out of it a pain, but truth be told the pickup wasn't much easier. The van's driving position was brilliant, and the short hood meant we could see into the cars ahead. (If a Sprinter van honks at you when the light turns green, it's because they can see you're on your phone.) I was rather surprised by the Sprinter's build quality, and not in a good way. Shortly after the van's arrival, I saw steam coming from under the hood, which turned out to be an improperly fit coolant pipe, which was fixed under warranty. When the parking brake stopped holding, the dealer diagnosed leaky rear axle seals that had allowed oil to seep into the mechanism (a drum setup separate from the rear service brakes) and contaminate the shoes. The fix involved replacing not just the parking brake but also most of the rear axle's components. Just before that trip to the dealer, the Sprinter threw a check engine light, which turned out to be from an improperly routed wire that had chafed and shorted. The passenger door was misaligned, but having it corrected would have required a trip to the body shop, and having been without the Sprinter for well over a week, I couldn't spare it any longer—after all, the mules need their hay. I left the door as-is. My primary reason for visiting the dealer was to troubleshoot our trailer brake controller installation. Turns out the van needed a software update to accommodate trailers with LED lights, which draw less power than incandescent bulbs and, in the case of the Sprinter, not enough for the van to detect the trailer. Trailers have been using LED bulbs for, what, 15 years? 20? You'd think this software would be standard by now. The one reason we didn't visit the dealer was for routine maintenance, because we only ran the van up to 16,358 miles, well short of the Sprinter's 20,000-mile oil change interval. Credit the Sprinter's commercial focus; a van can't make money when it's in the shop, so the goal is to keep it away. The Sprinter's engine holds 10.6 quarts of oil, so there's plenty of capacity for dirt. Even so, the Sprinter did ask for a quart of oil at around 12,000 miles. Yes, it asked; the Sprinter's engine has no dipstick. You check the oil through the dash display, and it lets you know when it's down a quart. We called a couple of dealerships to ask about the cost of the 20,000-mile 'A' service, which includes an oil and filter change, fuel filter change, DEF top-off, and general inspection. One dealer quoted $850, another said 'anywhere between $300 and $1,400' depending on what the van needed. The only other fluid we added was diesel exhaust fluid, a total of 12.5 gallons, which cost us $133.75. That averages out to 1,132 miles per gallon of DEF, just over half the consumption of our 2024 Chevrolet Silverado ZR2 diesel, which used a gallon of DEF every 579 miles. (That said, DEF usage does go up with higher loads, and we did more towing with the Silverado.) Like oil, the Sprinter asks for DEF when it needs it, starting when the underhood tank is 2.5 gallons low. DEF is sold in 2.5-gallon jugs, so that worked out perfectly. So Can a Van Replace a Pickup Truck, or What? A year of using the Mercedes Sprinter as my pickup truck has made me a van convert. With the notable exception of towing, everything our pickup can do, our Sprinter van did as well or better, including hauling, daily-driving, and off-roading. Yes, I would have liked a more upscale interior, and a year without drive-throughs did not have the intended effect on my waistline I was hoping for. A little more acceleration would have been nice, too. Of course, towing is a critical part of the equation for my own use case, and if we owned a Sprinter, we'd have to invest in a lighter-weight European-style horse trailer. (Most North American trailers have a tack and storage room at the front. European trailers dispense with the tack room, because if you're towing with a Sprinter van, you already have a storage room—the van. It's worth nothing that even with a 5,000-pound trailer, the Sprinter maintains most of its payload capacity.) Still, there are other vans that can tow more. Properly equipped, a Ford Transit can tow up to 7,500 pounds, and the aging Chevrolet Express can tow a trucklike 10,000 pounds. So my one major complaint about our van can be solved. Would I actually invest my own money in a van? Our old Chevy pickup still has plenty of life left in it, but come the day it needs to be put out to pasture, don't be surprised if you see a van parked in its place. For More on Our Long-Term 2024 Mercedes-Benz Sprinter: Our New Pickup Truck Is... A Van? Off-Roading in the Sprinter—There Is No Turning Back This Van Is a Pretty Great Pickup Truck Our AWD Sprinter is Kind of a Pain in the Ass A Mission From Dog Trial by Fire Towing with the Sprinter: It's No Mule Can Different Fuel Make a Sprinter Van Faster? Would You Rather Daily Drive a Huge Pickup or a Sprinter Van?
Yahoo
33 minutes ago
- Yahoo
Is This Stock the Best Way to Play Chinese AI?
While restricted from the latest and greatest chips for now, over half the world's AI researchers are in China. As such, growth investors should try to have some China-related AI exposure in their portfolios. Here's how a South African stock may offer the best way to play Chinese AI. 10 stocks we like better than Naspers › Would you believe that the best way to play China's AI ambitions may be a South African stock? Artificial intelligence has the potential to change the economy. As such, investors should have exposure to some AI companies in their portfolios -- and that includes Chinese exposure. Even Nvidia CEO Jensen Huang recently noted that China is home to half of the world's AI researchers. And the January unveiling of the groundbreaking DeepSeek R1 model showed how China is able to innovate cutting-edge AI, even with limited silicon access. Nevertheless, investing in China carries risks pertaining to its economy, its government, and the market mechanics of owning Chinese stocks. That's why the best way to expose oneself to Chinese AI may be South African holding company Naspers (OTC: NPSNY). It's still very early in the AI races, and in truth, it's hard to know who is going to win out. It could be an upstart AI lab such as DeepSeek in China or OpenAI in the U.S., or one of the existing big tech giants. In this investor's opinion, the existing incumbents are the best bet. These companies have huge technological and financial resources to invest in AI, and AI should also enable existing large businesses to boost revenue and lower costs through better automation and data science. That's why Tencent (OTC: TCEHY) seems like good bet to be a big winner from Chinese AI. Tencent has the largest social media platform in China in WeChat, with more than 1.4 billion users. It's also the largest video game publisher in the China, the largest video and music streaming company, one of the two large digital payments companies, and one of China's cloud computing giants. In addition, Tencent is building its own large model, called Hunyuan, which could lead to vast new AI opportunities if it ends being more performant than DeepSeek and other competitors over time. Tencent stock doesn't trade on U.S. stock exchanges, but it does trade over the counter as an American depositary receipt. However, a better way to play Tencent may be through its largest shareholder, European-domiciled Prosus (OTC: PROSY), which owns more than 23% of Tencent's stock. And the best way to invest in Prosus may be in its largest shareholder, Naspers, which owns 43% of Prosus. Naspers is a 110-year-old South African media company that was fortunate enough to invest in Tencent when it was just a startup in 2001. That investment eventually grew to be worth hundreds of billions, dwarfing Naspers' original business. Naspers then sold portions of its Tencent stake over time, investing in new businesses across food delivery, fintech, classifieds, e-commerce, ed-tech, and other tech startups. However, Naspers began to trade at a huge discount to the value of not only its net assets but also its Tencent stake alone. Management attributed that growing discount to its large size relative to the Johannesburg Stock Exchange. The thinking was that large index funds cap their asset weightings, which meant they had to sell Naspers stock once it reached a certain-sized allocation. In 2019, Naspers created Prosus, a new entity it listed on the larger Euronext Exchange in Amsterdam, and spun it off to shareholders. Today, Naspers owns about 43% of Prosus, which owns all of Naspers' assets outside of South Africa, while Naspers owns some small additional South African assets on its own. The plan didn't really work, though, as, Prosus still trades at a 30% discount its total net assets and even a 9% discount to the value of its Tencent stake alone. Meanwhile, Naspers trades at an even bigger 36% discount to the value of its assets, which is mainly just its 43% stake in Prosus. Owning Tencent at a discount is an attractive proposition, but of course it won't make much of a difference if the valuation gap never closes. That being said, a new development at Prosus could change this dynamic. Before this year, Prosus' businesses were free cash flow negative, so the company was dependent on Tencent stock sales and Tencent's dividend for cash for funding. However, in fiscal 2025, which ended in March, Prosus' non-Tencent businesses turned free cash flow positive for the first time, excluding Tencent's dividend, generating $36 million in positive free cash flow and improving free cash flow by nearly $1 billion over the past three years. Prosus also has a new CEO in Fabricio Bloisi, who took over the job exactly one year ago. Bloisi is an entrepreneur and the very successful former CEO of iFood, now the largest food delivery company in Brazil, which Prosus fully acquired back in 2022. Having a proven startup operator is a contrast with Prosus' prior leadership, and recent results appear to show Bloisi is raising the bar on execution. Before today, if Prosus wanted to buy a company or repurchase its stock at a big discount, it would have to sell part of its Tencent stake. That probably played into the "conglomerate discount," as investors may not have liked the capital allocation of selling a high-quality business in Tencent to buy more risky and unprofitable startups. However, if the other operating businesses continue to grow positive free cash flow outside Tencent, that could change sentiment. Prosus wouldn't have to sell as much or any Tencent stock and could benefit from future compound earnings growth. If a Prosus/Naspers valuation rerating occurs, that could make Prosus and/or Naspers the most attractive way to play Tencent -- and with it, the growth of Chinese AI. Before you buy stock in Naspers, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and Naspers wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $713,547!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $966,931!* Now, it's worth noting Stock Advisor's total average return is 1,062% — a market-crushing outperformance compared to 177% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of June 23, 2025 Billy Duberstein and/or his clients have positions in Naspers and Prosus. The Motley Fool has positions in and recommends Nvidia and Tencent. The Motley Fool has a disclosure policy. Is This Stock the Best Way to Play Chinese AI? was originally published by The Motley Fool
