Five tips for Cutting through the Hype and Avoid Pitfalls

The hype and feeling that everything is moving incredibly fast with AI is real. Everyone, more or less, knows about the existence of AI. Individuals or business executives often ask me about AI’s implications and opportunities. There is a lot of hype and new material about AI, which might intimidate anyone navigating this space. You shouldn’t. This article will share a few suggestions for navigating this busy, fast-moving space.

Author + DALL-E 3

We usually hear or read about concerns about AI taking over jobs or impacting the job market. However, I was recently asked, “Will AI replace my business?”. The answer is Yes if your business is brilliant but not defensible. This is often the case in the tech industry, which is evident because tech is fast-moving. If you followed the evolution of iOS and Android, you probably noticed how the early versions of these mobile operating systems were relatively bare. Over the years, several practical third-party applications such as Wi-Fi hotspot apps, torch/LED apps, data monitoring and battery-saving apps started to emerge. These were brilliant and added value. However, they were not defensible. They ended up being imitated and implemented into operating systems. Each was a good business concept, momentarily, that was not sustainable enough to survive long term.

Another realistic threat is that a competing business leveraging AI could give you a hard time. This would push any competitive entrepreneur to venture into this space and get consumed by the fast developments. This is normal, and it’s okay. The feeling is primarily due to the hype around AI. You only need to develop an intuition for handling change and innovating through it. Because of the hype, it’s understandable that executives feel lost or helpless. But this can change, starting today.

1) Understand the AI Hype

Just 80 years ago, the neuron was modelled mathematically for the first time, paving the way for artificial intelligence (AI) and artificial neural networks. While interest in AI gradually grew over the decades, with a few spikes and winters along the way, it remained mostly confined to scientists and engineers. They were those creating and using the technology. The skills gap needed to lessen for anyone else to participate. AI couldn’t scale to widespread use and adoption.

Google Trends results for the term “AI” globally for the past three years.

However, in 2022, we’ve seen a massive spike in interest, increasing by orders of magnitude. Generative AI models, particularly ChatGPT and other chat-driven apps, have made AI exceptionally accessible. Now, anyone with a mobile phone or laptop can interact with and use AI systems with just a few clicks from an easy-to-use website or mobile app.

Gartner’s Hype Cycle

Hype might have a negative connotation to it. It has, in its way, a significant impact on the lifecycle of concepts and ideas. My favourite source of well-informed insights about technological hype is Gartner. Among other material, they publish an annual Hype Cycle and have recently been dedicating one to AI. The hype cycle is a curve that merges a peaking hype on the left-hand side, gently blending with a lifecycle curve. Along the curve, one finds different technologies depending on how they are perceived. The technology closer to the left is an early innovation, susceptible to hype and inflated expectations. This is technology to keep an eye on, but it would carry a fair amount of risk. On the other end of the curve (right), we find technology whose interest plateaued into a stable adoption. This technology is (or very close) to widespread adoption.

Gartner’s AI Hype Cycle published in July 2023 (Source: Gartner)

2) Clear a Common Misconception

People give human-like characteristics and attribute emotions to complex non-human objects. AI is one of them.

This tendency is known as anthropomorphism and is a cognitive tool we use to handle and deal with complex systems. Attributing human characteristics to objects or concepts makes them more familiar and comprehensible. Simplifying complex systems or objects can easily lead individuals to refer to systems such as AI using pronouns. This would follow the line of thought of a metaphor that, while it serves to hide complexity, would still have gaps in its narrative and might limit people from understanding the limitations or challenges of this technology.

“AI As a general purpose tool” Dylan × DALL·E

The media also plays a role in this misconception. Movies, TV shows, and literature often depict AI and robots with human-like personalities, emotions, and intentions. These representations can influence public perceptions and lead to anthropomorphic beliefs about real-world AI, which is far different from what is depicted. Unfortunately, this can also lead to fear and uncertainty. AI is a transformative technology, and its rapid development can be unsettling. By anthropomorphising it, people might try to make sense of their feelings of unease or concern about its implications.

AI systems can perform tasks that were once exclusively the domain of human cognition, such as language processing or image understanding. This overlap of capabilities can lead people to mistakenly attribute human-like thought processes to AI. This is also reinforced by good user experience design, such as the excellent quality of interaction in systems like ChatGPT, Bard and Claude. Some people might also develop emotional attachments to technology, especially if it plays a significant role. For instance, someone might grow fond of their AI-powered virtual assistant because it provides company or assistance.

All of this is understandable. These ways of looking at complex systems can benefit individuals to deal with the ever-increasing complexity of AI. But AI is not a being — it is a general-purpose technology. Anthropomorphism can easily mislead our professional judgement when adopting or applying AI in an organisation to a problem.

3) Treat AI Like Any Other (Technology) Project

AI projects are no different than software projects or any other technology initiative. We can even abstract it further and consider an analogy of installing a new bathroom.

“A bathroom fitting that reminds you of tech” Dylan × DALL·E

You need fittings, water supply and infrastructure to deliver it to the fittings. With AI, the fittings are the predictive features — recommendation systems, analytics, object recognition, anomaly detection, and chatbots. Data is the water — it needs good supply and quality. Infrastructure delivers it — cloud, IoT, phones, websites.

Just like any project, focus is critical. You wouldn’t want to spend your entire bathroom renovation budget on fittings while ignoring water, plumbing and installation services.

4) Choose a Definition that you can Work With

As AI hype grows, so do definitions. There is no universal definition of AI. AI was initially defined in the 1950s by Stanford’s John McCarthy as “the science and engineering of creating intelligent machines”. The context of this definition was the discovery of applications of the back-then new electrical computing machine that was capable of performing tasks faster than humans in a more consistent manner. Understandably, its performance was attributed to intelligence. As decades passed and technology evolved significantly in all its forms, the term AI survived, but its context kept changing.

There are plenty of sensational definitions that compare AI to human intelligence. For decades, AI was traditionally seen as the venture towards creating programs that mimic human intelligence. Such visions can be healthy for the progression of science and engineering but risk being dramatised and misleading.

One can use a practical definition of machine learning, which involves programs that learn from data and improve with the more data they process. Rather than defining AI, I prefer taking a perspective of seeing it as the science of approximation. We use it where no engineering solution exists, training models to approximate tasks.

A restaurant can train an AI model to assess the size of its serving portions. It will improve customer satisfaction, quality control, and cost control while reducing food waste. (Inspiration from this BBC Programme)

From my experience in applying AI, definitions get in the way, and the perspective of using AI to approximate patterns in data leads to creative and valuable solutions. For example, unsupervised learning can be seen as a selection of tools that allow us to find and approximate trends in unlabelled data. This can be

5) Make the best out of Democratised AI

Above all, generative AI and its easy accessibility are democratising AI. Developers can experiment with more efficient code. If pitching ideas, you have a new tool for analysis and imagination.

Having the right tools is not enough (Image from here)

But challenges remain. A significant skills gap persists between conversing with a chatbot and deploying AI in an organization. Sustainable economic models based on AI require courageous educational reforms for prosperity.


As we look towards 2024, it’s healthy to cut through the noise surrounding AI and focus on pragmatic action.

First, get a handle on the hype by understanding where AI technologies are in their lifecycle; this will give you a realistic view of their maturity and business applicability. Second, resist the temptation to humanise AI; remember, it is a tool that can offer solutions, not a sentient being. Remember, it is designed to ‘sound’ human, but only for usability.

Similarly, manage your AI initiatives like any other business project, allocating resources wisely and setting achievable milestones. Don’t get trapped by debates over definitions. They are only suitable for entertainment purposes. Focus on how AI can solve your specific challenges.

Lastly, capitalise on the accessibility of AI tools now more than ever. The landscape is set for innovation, and the barriers to entry are diminishing. Will you observe the AI transformation or actively participate and steer your initiatives towards future success?

Pragmatic Ideas for Navigating the AI Landscape was originally published in UX Planet on Medium, where people are continuing the conversation by highlighting and responding to this story.