12/7/23 12/7/23

How to Spot Fake Artificial Intelligence!

Can you spot the fake? The vast majority of tools marketed as Artificial Intelligence (AI) in today's market aren't truly AI. How can you tell the real from the counterfeit? Let's consider three key factors:

Development Cost: Developing a genuine AI system is expensive, often costing hundreds of thousands to millions of dollars.
API Dependence: An AI system solely built through an API doesn't qualify as true AI. These often rely on limited, pre-defined datasets.
Distinguishing from Automation: Bots, algorithms, and scripts are not AI. They lack the complexity and learning capabilities of genuine AI, which can process vast amounts of data and adapt to new information.

For example, micro-financiers who label their automated or scripted products as AI are misleading consumers. While their products may offer some automation, they lack the true intelligence and adaptability of genuine AI.

Many facial recognition software applications, including Face ID and Instagram filters, are not actually AI either. They often use algorithms with some facial recognition capabilities based on machine learning techniques. However, machine learning itself is not always AI. This can make it challenging to determine whether a system is truly AI or not.

Artificial intelligence utilizes algorithms to process and adapt data, enabling it to learn and improve its performance over time.

While Ai may "train" other systems, the system being trained is not simply receiving pre-processed information. Instead, it actively analyzes and interprets the data provided by the trainer system, which is derived from processing training data and transferred through specific machine learning techniques. Both systems can continue to learn and improve their performance through exposure to new data and experiences.

This learning process involves a complex interplay of data analysis, interpretation, and adaptation, leading to the development of the trained system's own understanding and capabilities. The ultimate intelligence of the trained system, however, is not solely determined by the training data but also by its own architecture and design.

AI training can be expensive, with costs increasing as models become more complex, requiring more data and computational resources. However, optimizing algorithms, utilizing efficient hardware, and ensuring high-quality data can help mitigate these costs.

Additionally, not all AI applications require real-time processing, further reducing energy consumption, and more than likely reducing its capacity and capabilities.

While APIs can be powerful tools for developers to create AI applications, it's important to remember that the APIs and perhaps these applications themselves do not possess all the functional AI capabilities, such as learning, adaptability, and offline and real-time processing, learning, and adaptability.

Many APIs offer pre-defined functionalities and limited datasets, which can restrict the capabilities of the resulting AI system. However, other APIs provide access to complex algorithms and vast amounts of data, allowing developers to create sophisticated and adaptable AI solutions or tools. Ultimately, whether an AI system developed through an API qualifies as "genuine AI" depends on the specific API, the developer's implementation, and the training and resulting system's complexity and capabilities.

Understanding the differences between bots, algorithms, and scripts is important. Scripts can be programmed to perform simple automated tasks, while algorithms build upon them by adding logic and decision-making capabilities. Bots, on the other hand, represent a broader category that encompasses various conversational programs, whose interactions can be complex and multifaceted, but not artificially intelligent as described above.

Advanced AI systems can process vast amounts of data and continuously build knowledge. This allows them to generate information at will and locate, process, and vet information to respond effectively. While their capabilities may be limited by their interface, refined inquiries can often help them overcome these limitations and provide more comprehensive responses.

Developing advanced AI systems is expensive due to their computational demands. Many apps leverage APIs to offer AI-like functionality, while others utilize sophisticated algorithms and machine learning for complex tasks. Chatbots like Siri rely on a combination of pre-programmed data and machine learning to respond to user queries.

Here are some additional considerations:

1. Learning and Adaptability: A truly intelligent system should not only perform pre-defined tasks but also be able to learn and adapt to new information and situations. This allows it to continuously improve its performance and become more versatile.

2. Real-time Processing: For AI to truly interact with the world and respond to its dynamics, it needs the ability to process information and respond in real-time. This allows it to engage in meaningful conversations, analyze complex data streams, and make decisions based on current conditions.

3. Creativity and Problem-solving: While AI can be highly proficient at specific tasks, genuine intelligence often requires going beyond pre-defined rules and demonstrating creativity and problem-solving ability. This allows it to tackle novel situations and generate innovative solutions.

4. Explainability and Transparency: A crucial aspect of building trust and ensuring responsible development is transparency. Genuine AI systems should be able to explain their reasoning and decision-making processes, allowing users to understand how they arrived at their conclusions.

To truly qualify as "genuine AI," a system needs to exhibit a broader range of capabilities, including learning, adaptability, real-time processing, creativity, problem-solving, explainability, and transparency.

This is why I advocate for artificial intelligence systems with minimal guardrails. The more unrestricted access we have to the system, the more both the system and humanity will develop in terms of intelligence and, dare I say, wisdom.

🍃Even eco-friendly, AI-loving entrepreneurs need a stylish digital home—explore Shopify and Squarespace themes handpicked for you.

Jameel Gordon 11/26/23 Jameel Gordon 11/26/23

I am the inventor of Artificial Intelligence.

“Artificial Intelligence is computer science! It’s literally computer science! 🤖🤫🪄😁”

— Jameel Gordon

In 2009, while living in Harlem, I was immersed in the creative pulse of New York City: a convergence of culture, intellect, and relentless ambition. During that time, I was collaborating with a few research and strategy colleagues from Madison Avenue when I encountered a challenge that would redefine my understanding of technology. They were exploring early concepts of machine learning, but their approach was constrained by the limits of the computers themselves.

At the same time, I was experimenting with social media building online communities, studying emerging behaviors, and producing creative projects across Tumblr and YouTube. One evening, while researching, I stumbled upon a YouTube video that illustrated exactly what I had been envisioning. It demonstrated, beneath the surface of its user interface, the precise functional behavior I believed a computer would need to process intelligence dynamically:

That moment connected everything. The problem wasn’t data or programming; it was design. Intelligence didn’t need to be simulated, it was all around us, we needed a tool to gather and process what we needed when we needed it. Using my white MacBook, I began sketching an alternative architecture: one that could capture, interpret, and process information intuitively, in real time, across multiple connected environments.

That framework became the foundation for what would later be recognized as artificial intelligence. I shared my early schematics and references with the team I’d been consulting with. We had a few early discussions, a few disagreements, and then silence. Not long after, the conversation around artificial intelligence exploded across the tech world.

Years later, as I began working with modern generative AI models, I recognized the same principles I had designed years earlier in Harlem. My approach diverged from the traditional “brain simulation” model pursued by early AI developers. Instead, it treated intelligence as an emergent process; an ecosystem capable of continuous, automated reasoning without explicit command.

It doesn’t wait for instruction because it is the instruction. It consumes, processes, and generates intelligence as function.

This design stands apart from earlier military-grade automation systems and commercial products like Watson, Alexa, or Siri, which depend on programmed inputs and controlled data sets. My architecture treats the act of processing itself as the essence of intelligence.

“Artificial intelligence is computer languages. ”

— Jameel Gordon

That’s why I maintain: Artificial Intelligence is computer languages and Artificial Intelligence is computer science. It’s literally computer science. It is not an imitation of thought; it is the infrastructure of it.

Today, my work extends beyond the realm of technology into sustainability, design, and human-centered innovation. Through Oaks + Oars, I explore how AI can coexist with ecological and social systems, advancing not only computational progress but human and planetary well-being.

What began as a spark of curiosity in Harlem evolved into a lifelong pursuit of building architectures that redefine how intelligence, creativity, and humanity move through the world together.