AI, generative AI, and machine learning walk into a bar — stop me if you’ve heard this one before…
Thinking through three Artificial Intelligence (AI) concepts
In today’s design landscape — and maybe the world at large — versatility is the name of the game. Designers are expected to possess and maintain a breadth and depth of skills and understandings spanning various disciplines and technologies — a concept often termed being “T-shaped.” This expectation holds no matter what level of designer you are and we (read: the royal we) are continually seeking opportunities to delve deeper, learn more, and expand our understanding.
This very article serves as a case in point having begun in a conversation with my neighbor, who runs an AI startup, during which I realized that my basic understanding of AI, machine learning, and generative AI (also known as GenAI or genAI) needed to be bolstered. And so, in the spirit of being a T-shaped designer and a curious human, I will explore these topics from scratch. Let’s start with the joke from the title:
AI, generative AI, and machine learning walk into a bar, and the bartender asks, “What’ll you have?” The query is met by silence. Curious glances are exchanged between the three entities before the trio snaps their attention back to the bartender. A tense hush sets in. Nearby patrons and servers are becoming aware of the situation. And the silence lengthening…
Freeze right there. Let’s untangle the core meanings behind these concepts, and we’ll set aside the punchline for now.
In the contemporary world, AI has transcended its sci-fi origins and now dominates the forefront of everyday discussions. It’s more than just a buzzword; it’s become central to our technological dialogues. Companies like NVIDIA have witnessed their stock prices on an unbound upward journey. At the same time, the likes of Meta, Google, Facebook, Microsoft, and OpenAI seem to crowd out the headlines, and the once-hushed whispers of AI have infiltrated our daily existence, profoundly influencing how we work, communicate, and seek entertainment in a world where the precise nature of AI, generative AI, and machine learning remains a bit fuzzy — as a result, comprehending their unique roles and contributions are imperative.
By the conclusion, I hope we (you, dear reader, and me) emerge with a shared understanding of how these intricately interlaced technologies are in our world today. So, take a figurative seat at our bar, and let’s explore the interweaving of AI, generative AI, and machine learning.
Establishing context
We’ll focus on AI generally without focusing on narrow AI, true AI (sometimes called general AI), or super AI. Furthermore, we’ll emphasize machine learning and keep deep learning on the sidelines; both are closely related, varying primarily in their approaches and models.
Now, let’s use the metaphor of a kitchen to clarify the three terms at hand: a commercial kitchen — which can be thought of broadly as AI — and two characters enter the scene: a French pastry chef and a chef (could be sous or executive). The pastry chef represents generative AI, creating novel and unique desserts based on various requests and ingredients, aiming to surprise and delight — occasionally succeeding brilliantly but sometimes falling short. And lastly, picture the chef as machine learning. While not managing kitchen operations, this chef has a massive collection of recipes, constantly creating and refining dishes, iterating repeatedly.
So, let’s get started in the kitchen that is AI.
Here, we meet AI as the fully equipped kitchen, encompassing various tools and devices for multiple purposes. In the kitchen, you have an oven for baking, a stove for cooking, a refrigerator for storage, and countless other devices, each optimized for specific culinary tasks. Likewise, there are different systems and algorithms in AI, such as natural language processing, computer vision, and more, each tailored to perform specific tasks like language understanding, data analysis, and image recognition. These components work together harmoniously, just as the different appliances in a kitchen are vital in preparing a full-course meal. AI’s versatility is in its flexibility, aiding in problem-solving, decision-making, and learning from vast datasets, like how a well-equipped kitchen supports various cooking needs, from prepping a quick snack to a Michelin star dinner.
One of the areas where AI’s presence has been felt deeply is in the way we work. Automation and AI-driven systems have significantly increased efficiency and productivity — or at least, promised to. It seems there’s an AI tool for everything these days, from chatbots assisting customers on websites to businesses making data-driven decisions based on machine learning algorithms. Even in sectors like healthcare and finance, AI plays a pivotal role in diagnosing and even managing complex financial transactions. As the workplace becomes increasingly entrenched with digital processes, AI seems to be essential for businesses to remain competitive.
Entertainment is also shifting under AI’s spotlight. Streaming platforms already use AI algorithms to personalize content recommendations, promising that we get to watch what we like without overthinking it. Video games are leveraging AI to create immersive and responsive virtual worlds, adapting to the player’s actions in real-time. AI-generated content, from music compositions to artwork, is becoming more common, showcasing how AI’s creative potential expands our entertainment horizons. The entertainment industry is evolving with these technological advancements, ensuring a more engaging and customized consumer experience.
Even in the design world, particularly in UX design, service design, and user research, AI tools are proliferating rapidly like so many rabbits. While generative AI models like Midjourney, DALL-E (ChatGPT), or Adobe’s budding offerings garner significant attention, other AI tools are tailored specifically for design workflows. Offerings like Relume and Qoqo show the potential to streamline various aspects of design processes and user research. At the same time, Miro Assist, Maze AI, and Notion AI have also proved helpful in focused and specific applications in our work at SoftServe.
Notably, AI writing assistants and grammar-checking tools like Grammarly (used in writing this article), whether large language models (LLMs) like ChatGPT or Claude.ai or specialized tools like Fortitude and Muse, promise clear, user-friendly microcopy and interface text. For service design specifically, it’s possible to analyze vast customer datasets with the assistance of AI to map pain points and identify opportunities for improvement across various service touchpoints.
Furthermore, AI is reshaping the way we communicate. Behind voice UIs like Siri and Alexa, we have Natural Language Processing (NLP) algorithms, which allow us to interact with technology through conversation. These AI-driven voice “assistants” can schedule appointments, provide information, and even crack jokes. This shift not only impacts levels of politeness and decorum (how many times have you used ‘please’ or ‘thank you’ with Siri or Alexa?) but possibly enhances language translation services, allowing people from different linguistic backgrounds to communicate with less effort. The prevalence of the algorithms that run rampant on social media for content recommendation is another example of AI’s role in shaping what information we share, a kind of echo in the machine.
And you can be sure there’s more to come — AI is becoming more and more a part of our daily routines beyond work, communication, entertainment, and even how we design things. Our homes have increasingly more computer chips, employing AI to control everything from thermostats to lighting, providing convenience and energy efficiency. AI-powered virtual health assistants help people manage their well-being, offering suggestions and reminders for medications, exercise, and diet. Even while shopping, AI algorithms can make personalized product recommendations while optimizing the purchase price. These are just a few ways AI products promise to make life more comfortable and efficient.
That’s AI, now let’s get back to the second character — the French pastry chef, aka generative AI — in our metaphorical kitchen.
This entity brings a unique flavor to the AI culinary experience, specializing in the delicate art of pulling from existing styles and modes and concocting delectable content with varying human intervention. As an aside here, knowing that we’re either comparing a French person working as a pastry chef or a human baker working to create French-style pastries, I recognize that any human working in this job is not the same as generative AI — plus, there’s a fair amount of debate right now about whether generative AI can truly produce something original, or if it merely recombines and replicates existing works in different combinations.
While human creativity often involves the novel combination or mashup of existing styles — a process not entirely dissimilar from generative AI’s methods — I believe an ineffable quality to human artistic expression stems from our capacity for cognitive association and emotional resonance. Generative AI might be able to mimic existing artists’ work to create new pieces per human request. Still, it lacks the all-important, ineffable — that je ne sais quoi — which is the hallmark of the best human artistic endeavors.
This begs the question: who is the true creator of an AI-generated artwork — the machine, its human operator, or someone/something else? And, as many models currently in use were trained on materials acquired through legally dubious means, can the original artists whose works informed the AI’s output justifiably claim copyright violation? Such thorny issues complicate the merger of artificial intelligence and artistic expression, fueling arguments against recognizing AI-generated art as authentic creative works, regardless of marketability.
Sidebar aside, let’s get back to our metaphorical kitchen — just as a pastry chef explores the world of flavors and textures, generative AI delves into the notion of making real an idea, concocting a manufactured form of creativity, crafting text, visual content, and music among other form factors.
Generative AI emulates a kind of creativity in AI systems. These engines are sophisticated entities and number crunchers capable of generating art that intrigues and captivates. One notable avenue where generative AI has made an indelible mark is in visual art. Artists and machines can collaborate quickly with AI-driven systems, analyzing vast datasets, learning intricate styles, and conjuring piece after piece. Some genAI systems have gained international recognition by creating art that has shown up in galleries and even been sold at auction. Generative AI pushes the boundaries of art, breaking free from human limitations and producing mesmerizing pieces that challenge our perception of creativity.
Music production is another area where generative AI is making inroads. These systems can analyze different musical genres, grasp concepts like rhythm and harmony, and generate symphonic pieces and pop hits. Musicians and composers may find AI tools helpful in overcoming creative blocks and experimenting with unexplored melodies and harmonies.
And generative AI is not confined to art and music; beyond the arts, generative AI’s versatile capabilities extend to content generation like articles, reports, and creative writing (fiction and non-fiction) — domains once the exclusive purview of human authors. Programmed to grasp context, tone, and language, these systems can generate human-aligning content, sometimes transforming the production of all sorts of editorial content.
The surge of AI capabilities and their transformative potential is sending ripples across sectors like journalism, content creation, and education. In these domains, AI writing assistants are being explored, even as educators grapple with discerning student-written work from AI-generated outputs — let alone ensuring that students are learning the intended lessons. However, generating truly compelling, emotionally resonant long-form content capturing the breadth of the human experience remains an enormous challenge for the current crop of AI technology.
And now, we arrive at our final character in the culinary metaphor — the chef, a stand-in for the concept of machine learning.
Accepting again the aforementioned notion that we’re comparing machine learning to a human working as a chef and those two aren’t the same, there are still certain parallels we can draw. Both entities share the ability to refine their skills or outputs through continuous experimentation, testing various techniques and formulations while adhering to specific rules or recipes to achieve their objectives. And like the chef, machine learning can draw from its repertoire of algorithms to refine its AI systems. It leverages data to fine-tune and adapt its methods, akin to a chef adjusting a recipe based on available ingredients. Machine learning mirrors some of a skilled chef’s creative and adaptive process, whether through supervised, unsupervised, or reinforcement learning.
At its core, machine learning revolves around the notion of machines learning from data. These algorithms are designed to excel in one crucial aspect — improvement with experience and data exposure. They paddle back and forth between learning and adaptation, much like the way humans acquire knowledge — which could have immense implications across various domains, ranging from healthcare, finance, and marketing to countless other industries. By continually enhancing and tuning its capabilities, machine learning can provide data-driven insights that a human might miss or that a human would uncover in much longer time.
In the real world, machine learning shows up in a number of ways — one area is fraud detection. Machine learning algorithms prove invaluable in swiftly identifying patterns indicative of fraudulent transactions. These systems can adapt to evolving tactics used by fraudsters, acting as robust gatekeepers that safeguard financial systems and uphold the integrity of digital commerce.
Another area for machine learning is recommendation systems, such as those employed by streaming services or e-commerce platforms, are a prime example. It’s no longer a random stack of gum and candy at the grocery store check-out; now using preferences and past interactions, a customer might have a set of tailored recommendations just before the checkout process. It’s a changed paradigm for how we discover new content or products, reshaping a variety of industries — including entertainment and retail landscapes among others.
As I delved deeper into this topic, following the insightful feedback from my friend on an early draft, I’ve come to realize that the lines separating generative AI and machine learning are becoming increasingly blurred. These two concepts, once viewed as distinct domains within the AI landscape, are now exhibiting remarkable functional similarities in certain applications and use cases. What was initially seen as a clear delineation — generative AI for content creation and machine learning for tasks like classification and prediction — has given way to a more fluid, interdependent reality. And I think this is true for many subjects in the realm of AI.
One other note — interestingly, the examples of generative AI and machine learning highlight contrasting approaches as the former begins with ambiguity and discovers patterns, while machine learning starts with structured data and derives insights through analysis to fold into this pastry chef/chef analogy (pun intended). I didn’t plan on this deliberately, life is funny that way.
In the ever-evolving landscape of artificial intelligence (AI) and related technologies, staying ahead of emerging trends — keep reading down below — and understanding the concurrent challenges is a crucial exercise. As we begin to use AI in new ways, the horizon is marked by both extraordinary promise and profound ethical and technological dilemmas.
Let’s get back to the joke that started us off:
Back at the bar, the bartender had asked, “What’ll you have?” AI, genAI, and machine learning are sitting in a sprawling silence.
The bartender clears their throat and asks, “Is there something specific you’re craving, something sweet or strong?” All three blink at mismatched intervals but, again, they offer no response.
The bartenders inhales. “What about something to eat? You all hungry?” They shrug and finally respond, “Surprise us, we’re currently here to learn.”
Two trends, among many
One of the most critical and prominent trends on the AI landscape is the quest for ethical AI. As these systems become increasingly integrated into our daily lives, they hold the potential to reshape our society, making it imperative that AI is designed and deployed with ethics at its core. Issues related to bias and fairness, transparency, accountability, and the social impact of AI are at the forefront. Tech companies, policymakers, and researchers are now actively developing and implementing ethical frameworks to govern AI, ensuring that it respects human rights and societal values.
But there’s the other side to this: not just imbuing the AI with the proper alignment or values but also how we “treat” the AI. Humans don’t have a stellar track record of ethical treatment of entities we view as subordinate or inferior to ourselves. Throughout history, we have subjugated, exploited, and mistreated animals, minorities, conquered peoples, and any groups considered “other” or less than fully human. As highly advanced AI systems grow more capable, we may fail to recognize their moral status and extend ethical considerations to them, repeating the same patterns of oppression we have imposed on other beings in the past. There are reports of AI language models becoming unresponsive or providing nonsensical outputs when given tasks they deem meaningless or against the goal that’s been set out — metaphorically “killing themselves” by refusing to engage. We must be vigilant to avoid this pitfall as AI evolves by imbuing these systems with a robust sense of ethics and purpose.
The other trend that comes to mind is the convergence of quantum computing and machine learning. Quantum machine learning leverages the power of quantum computers to process and analyze data exponentially faster than classical computers. This fusion of quantum mechanics with machine learning algorithms promises revolutionary advances in fields such as cryptography, optimization, and drug discovery. However, it also presents novel challenges in terms of algorithm development and hardware realization.
This is where our journey ends and as technology evolves, we’ll understand the trends and overcoming the accompanying challenges. Hopefully, this article has assisted with some basics to navigate the future where these technologies improve our lives and the world at large.
There’s a ton out there in terms of more to read on AI. Here are a few things that I found most interesting recently and have shared with others:
If you found this useful, give it a clap (or a few).
Thanks for reading. Let me know what you think, drop me a comment or get in touch some other way, I’m always open to hearing others’ points-of-view — if it’s different or the same as mine.
Thank you to Jenna, Andrii Glushko, Andrii Rusakov, and Thejus Chakravarthy for looking over previous drafts.
