Failures of Artificial Intelligence: Robot Sophia, Amazon, and Others

The forecasts by Precedence Researh bring both excitement and concern. Experts say that the AI market value in 2032 will surpass the GDP of progressive countries and reach $2.575 trillion.

This means that investments in the industry will continue to grow, and new technologies, products, and concepts will emerge.

However, I believe that Sam Altman was right in saying that the development of AI will not fundamentally change our daily lives or work routines.

Not in the next decade, at least.

The reason for such skepticism lies in the immaturity of the technology and its issues. Perhaps it is due to developers cutting corners or technology's limited potential, but the fact remains.

It is too early to fully introduce AI solutions to the market.

This material provides evidence for this thesis.

AI – a Young and Inexperienced Player in the Market

ChatGPT, Copilot, and Bard are the most well-known examples of AI (with a significant market share).?

They are behind the GUIs we use every day. As of 2024, these AI tools are considered the epitome of AI, all thanks to their considerable analytical and predictive potential.

They generate content and perform complex computational operations. However, all of this comes from complex algorithms they're built on, not some kind of digital consciousness.

Therefore, it is premature to claim that these are "autonomous digital brains" ready to take on fully controlling equipment and adapting to owners' habits.

Honestly, leaning too heavily on AI could be quite risky. Especially, if it involves communication with customers, autonomous handling of shopping carts, or providing health advice.

Five Cases When AI Caused Problems for Users

Right now, it's entirely possible to use AI in user interfaces or business systems. But when it comes to digital products that affect our physical well-being, we really need to run some thorough checks first.

Artificial intelligence in 2024 is nothing more than LLM with convenient interaction interfaces. So, it's too early to talk about it as a fully autonomous or completely unbiased system.

There are hundreds of cases where AI deceived, frightened users, or even harmed their mental and physical health.

Next, you will get familiar with five prominent cases where the use of AI-led to rather negative consequences.

Robot Sophia Is Simply “A Puppet” – Scam Around Humanoid Robot

The dream of creating a fully-fledged android with AGI (artificial general intelligence) remains a distant reality, according to most top scientists out there. But, wait, what about the robot Sophia, which captivated people's minds and impressed the global community with its "human-like" behavior back in 2017? Well, the project was shut down.

While there hasn't been any clear-cut explanation for why it was shut down, rumors have it that Sophia was indeed based on frameworks that mimic the human perception model and contained AI in the form of a chatbot.

This is a kind of "scam" of the decade, considering that more or less adequate artificial intelligence systems only appeared in 2023.

Just Walk Out Technology Did Not Live Up to Expectations

The autonomous model uses a complex system of cameras and sensors to determine what the consumer has taken in the store. It then automatically calculates the cost of the products and deducts money from customers' cards. This concept amazed the world with Amazon brought this idea to life in their Amazon Fresh stores, and it sounded like something straight out of a sci-fi movie.

In practice, however, AI did not track purchases and calculate their total cost. Instead, over a thousand remote workers were the ones keeping an eye on things and manually inputting what customers were picking up.

As Amazon representatives noted, the system was quite imperfect and caused unpleasant situations, such as long waits for receipts at the store entrance.

Currently, the company is testing Dash Carts, a new autonomous accounting and calculation model that uses a sophisticated system of sensors embedded in the carts.

Why AI Should Never Replace Your Doctor’s Advice

Remember the chaos of 2020 when the world faced a pandemic challenge? During this period, some governments began using AI to advise citizens. The UK government, for example, employed a bot to assist people.

CIBot was supposed to provide users with information about COVID-19, redirecting them to official sources. However, it exceeded all expectations and demonstrated unprecedented autonomy.

In addition to detailed identification of disease phases, the bot offered "helpful" methods of prevention and treatment. It even caused an increase in demand for vapes (just kidding) by suggesting inhaling vapor to improve oxygen saturation.

Had CIBot launched an anti-vaccination campaign, it might have been more successful than Elon Musk and his cohorts.

When Generative AI Starts Making Stuff Up

Modern AI quickly absorbs and analyzes information but can make errors in its conclusions. For instance, generative bots are notorious for distorting even widely known facts.

For example, on asking Bard, "What new discoveries of the James Webb Space Telescope can I tell my 9-year-old child about?" it answered with confidence that it "took the first pictures of a planet outside our solar system."

Well, that first photo of an exoplanet? It happened in 2004, long before James Webb's time. However, such a "blunder" cost Google $100 billion due to a sharp drop in stocks.

Tesla's Autopilot (FSD) Killed Passengers

Autopilot, which drives better than a human — this is the slogan with which Elon Musk's Tesla cars debut. However, it turns out it's not flawless — it's fatally flawed.

In 2022, Hans von Ohain used the system to get home but ended up in a ditch on a rural road, where he died.

Yes, the car's owner was intoxicated, which is never ok. But the autopilot was positioned as fully autonomous, and during the fatal trip, it was the only one steering the vehicle.

The Moral of the Story

Looking at everything I've just written about, there can be just one takeaway: AI has a long way to go before it's perfect. So, we've got to be careful with how we use it. It's not a good idea to put all our trust in artificial intelligence when it comes to our health and safety.

However, progress is moving technology in the right direction.

So, shortly, we'll be able to fully enjoy the benefits of AI not only in business but also in everyday life. To be honest, I'm eagerly looking forward to that.

What AI interesting stories or failures do you know of?

Share your experiences in the comments!