The Commoditization of Knowledge - Part Two

In the first part of our series, we looked at how Generative AI is making knowledge both more accessible and easier to apply. In particular, we discussed how such a technology is changing how humans understand and interact with information. One need only look at GPT-4 and Google's new model, Gemini, to realize just how far AI has come in these capabilities.

At Wizeline, we believe that AI is not meant to replace human intelligence but rather to enhance it. This second part of our series will focus on the areas for which AI (still) needs human input and guidance, highlighting the importance of collaboration between humans and AI. It's all about working together, using AI's strengths to augment our own, and focusing on the unique skills and qualities we humans possess.?

The Next Frontiers of Intelligence?

Because of the huge progress in AI brought on by the introduction—and rapid development—of LLMs, it's clear that the landscape of intelligence is evolving. But this evolution raises an important question: in which areas does human intelligence still hold the upper hand over AI? This part of our series aims to explore some important human capabilities and compare them with their current counterparts in AI. We'll explore key domains such as learning, reasoning, creativity, and ethical judgment, among others, trying to carefully point out where AI excels and where the depth of human intelligence continues to be irreplaceable. This comparison will inform us about AI's current capabilities and help us envision how humans and AI can best collaborate.??

Let us break down the capabilities mentioned earlier as follows:?

Learning and Adaptability

Learning and Adaptability refer to acquiring knowledge or skills through experience and then applying that learning in new, unforeseen, or changing environments. With this definition, let's compare human learning with AI learning.

Human vs. AI Application

Humans: we learn from all our life experiences and can very quickly adapt to new situations. This adaptability is evident in our daily lives, where we can swiftly adjust to new information or environmental changes. Human learning is rich, contextual, and often nuanced. It involves a wide range of cognitive, emotional, and social factors. For instance, it is clear that we use our emotions to guide our decision-making and learning, and we rely on social interactions to learn from others and share our knowledge and experiences. Thus, our learning is not limited to single tasks but thrives on interconnections. Think of the fact that we can learn the alphabet using a song, or a different language through games, or how to keep rhythm by dancing.???

AI: ?while it's true that AI has grown significantly in this area (namely thanks to machine-learning algorithms that process large datasets and help us analyze new data), it's also true that, compared to human flexibility, machine adaptability is very limited when it comes to highly unpredictable or novel scenarios. AI tends to rely on the data it has been trained on and struggles with situations outside these parameters. Similarly, AI focuses on pure data analysis and cannot incorporate emotional or social factors into its learning process. In turn, recent research trends focus on making AI more adaptable and flexible, akin to human learning abilities.

Example: if we asked the reader to evoke personal memories from their life completely unrelated to AI or to this article, the reader will most likely be able to do so without any problem. We can ask the reader to tell us a personal learning experience. In turn, AI will definitely struggle with jumping to a completely different element from the one trained.??

Reasoning and Problem-Solving

Reasoning refers to a collection of cognitive processes that include analyzing information, drawing inferences, making judgments, and solving problems. It involves logical thinking, understanding contexts, and deducing consequences. Just as we did for learning, let's now compare humans with AI in this respect.?

Human vs. AI Application?

Humans: we excel in abstract, detailed reasoning, drawing conclusions from various events and all our knowledge (be it theoretical, practical, by association, or by acquaintance). We can understand implications, make judgments in complex scenarios, and think conceptually and creatively.?

AI: AI has advanced in specific, rule-based reasoning, performing well in structured environments. However, it struggles with abstract and context-driven reasoning. For instance, AI can be pretty good in some kinds of mathematical reasoning or code generation, but it may find navigating complex ethical dilemmas or intricate social contexts quite challenging.

Example: if we ask the reader to infer the next number in the series: 4, 7, 12, 21, they will have no problem figuring out that the next number is 38 since the pattern in the series is given as follows: 7=(4x2)-1, 12=(7x2)-2, 21=(12x2)-3, 38=(21x2)-4), etc. In turn, we encourage the reader to try out this exercise with any state-of-the-art text generator or with any series solver online to check for their reasoning capabilities.

Creativity?

Creativity involves generating original ideas or solutions. It encompasses logical thinking, innovation, and the ability to conceive new concepts or approaches. As for the comparison between human creativity and AI creativity, we can put forward the following points:

Human vs. AI Application

Humans: we are known for our (seemingly) boundless creativity. We often use abstract reasoning in doing so, taking inspiration from various sources such as influences, teachings, personal memories, and individual beliefs. Human creativity is not just about generating something new; it's about adding value, meaning, and depth, whether in arts, science, or everyday problem-solving.?

AI: it's true that it has begun demonstrating what some can call creativity, particularly in simple design tasks. While AI can generate new content, like a poem or music, it often lacks the depth, emotional connection, and originality inherent in human creativity. For instance, AI can create a new piece of music, but it may not evoke the same emotional resonance as a piece composed by a human. Furthermore, in this matter, the role of the spectator is key. AI can generate content, but it's most likely the spectator who interprets it as art.?

Example: while AI image generators are impressive , they cannot generate a drawing, painting, or image from, let's say, the reader's childhood. The reader definitely can, and even if the result is not a masterpiece (although we're not saying it can't be), it will surely include emotional content and character. The latter notions are still blatantly absent in AI-generated work. Another interesting point in this respect is that, while human creativity thrives on constraints , Gen AI? does not fare well with restrictions. For example, try to ask any text generator to write a poem with no "a" 's and check the result.

Agency?

Agency refers to the capacity to make choices and take actions autonomously, where such choices and actions are based on one's own volition, judgments, beliefs, desires, preferences, intentions, obligations, emotions, intuition, and perspectives on the situation at hand.?

Human vs. AI Application

Humans: our sense of agency is characterized by our decision-making ability. As implied by the definition of agency we gave above, humans can weigh a wide variety of epistemic, intentional, moral, ethical, and emotional factors in their decisions. Such a combination of factors helps us easily adapt to new and unforeseen scenarios.

AI: AI lacks true agency . We say this because AI's decisions and actions are bound by the programming and algorithms created by humans. While AI can appear to make decisions, these are the result of data-driven predictions or predefined rules. AI's decision-making process lacks the intuition, moral reasoning, and emotional consideration that humans possess.?

Example: when asked to reflect on their daily chain of actions, the reader will most likely know where these actions come from and why they were chosen. However, decisions and actions in most AI systems depend on decisions and actions their designers and users are making (think of the YouTube algorithm or Google Maps). AI systems can make suggestions for decision-making, but humans are the ones ultimately calling the shots.?

Goal-Directed Intentionality

Goal-Directed Intentionality involves setting objectives based on one's desires or needs and taking action to achieve those goals. It encompasses the capacity to formulate aspirations and strategically work towards them.

Human vs. AI Application:

Humans naturally set personal goals and aspirations driven by individual desires, needs, and values. Our goal-directed intentions reflect complex motivations and long-term planning.

AI: AI systems can work towards programmed goals, but human developers determine these goals. AI is not currently able to self-formulate intentions or aspirations. It follows predefined objectives and does not possess the intrinsic motivation found in humans. Current AI systems cannot autonomously set goals or intentions as humans do. They remain tools directed by human input and objectives.

Example: just as implied by the example for agency, the goals of current AI systems are aligned with human objectives. For instance, whenever we use an AI assistant (like Siri, Alexa, Google Maps, etc.), we have an underlying purpose that refers to our own activity (finding info, searching the web, seeking for locations). Even the most powerful Gen AI tools cannot formulate their own goals (yet). Once again, we encourage the reader to try and challenge any state-of-the-art text generator to reveal something similar to a true personal objective.?

Awareness

Awareness in this context refers to self-awareness and consciousness. It's the ability to be aware of oneself, one's actions, and the implications of those actions. It also involves understanding one's existence and environment. Comparing AI with humans in this respect, we can say the following:

Human vs. AI Application:

Humans naturally possess self-awareness and consciousness since we can grasp the nature of our actions, reflect on them, and foresee their consequences. This awareness also allows us to understand our place in the world and how we relate to others.?

AI: AI, as it currently stands, lacks self-awareness and consciousness . It is based on algorithms and has no true comprehension of its own processes or the implications of its actions. AI does not have a sense of self or consciousness like humans do. Its actions result from programmed responses and data-driven decisions, not from any form of conscious thought.?

Example: similar to the other experiments we have hereby proposed to the reader, it's interesting to compare what happens if you ask any human to provide a short, meaningful bio on themselves as opposed to giving the same task to a text generator. Try it out, and the differences concerning self-awareness will be apparent.??

Emotional Intelligence?

Emotional Intelligence is the ability to understand, interpret, and respond to emotions in oneself and others. It involves recognizing emotional cues, empathizing, and appropriately reacting to those emotions. As far as comparisons between AI and humans, consider the following points:

Human vs. AI Application:

Humans: our ability to read and respond to emotions is deeply ingrained in every sphere of our activity. We use a combination of verbal cues, facial expressions, body language, and situational context to understand and react to emotions. Our emotional responses can be refined over time, and they revolve around the key notion of empathy.?

AI: through 'emotion AI' or affective computing, AI can simulate responses to emotional cues. However, this does not mean that AI genuinely experiences emotions. Similar to what we have mentioned for most AI systems throughout this part of our series, AI can use data to recognize emotional patterns (like voice inflections or facial micro-expressions), but AI's treatment of emotions is confined to only these data patterns. Even if Gen AI can classify human emotions and generate responses to them, the majority of emotion AI is based on highly restricted science. Emotion AI algorithms (even when trained on diverse data sets) reduce facial and tonal expressions to an emotion without considering social and cultural backgrounds.?

Example: while algorithms can recognize that a person is crying, for instance, it is not always possible to accurately point out the reason behind the tears. Similarly, a scowling face doesn't necessarily imply an angry person, but that's the conclusion an algorithm will likely reach. Human's empathic abilities, in turn, can easily determine the reason behind those tears or whether the scowling face means genuine anger rather than acting or humor.?

Ethical and Moral Reasoning

Ethical and Moral Reasoning evaluates situations, actions, and outcomes based on moral and ethical principles. It relies on the notions of right, wrong, justice, fairness, and virtue.

Human vs. AI Application:

Humans: we constantly integrate moral and ethical beliefs into our choices, and it's a combination of cultural background, education, and societal norms that shapes this integration. For instance, when faced with a moral dilemma, a person might consider the impact of their actions on others, basing such considerations on principles like fairness and responsibility. Human ethical reasoning is complex as it presupposes both emotional empathy and the measuring of various long-term consequences.?

AI: in contrast, AI's approach to ethical and moral reasoning is confined to its programming. AI systems make decisions based on algorithms that incorporate ethical guidelines set by their developers. They lack innate moral understanding and cannot inherently discern right from wrong beyond their coded instructions.?

Example: an AI programmed for judicial decision-making would rely on encoded principles of law and ethics. Still, it would not understand all the moral implications of its decisions that a human judge would consider. AI's ethical reasoning is as advanced as the frameworks and data provided, and it requires human oversight for complex (or unprecedented) ethical scenarios.?

Augmented Human Intelligence?

With our exploration of the advancements and limitations of AI compared to human intelligence, it's evident that while AI is quite good in areas like learning, problem-solving, and simulating creativity and emotional responses. However,? it significantly lags in self-awareness, genuine creativity, ethical reasoning, and autonomous intentional decision-making. The quest for Artificial General Intelligence (AGI), an AI that mirrors the full range of human cognitive abilities, remains an ambitious and unrealized goal.

The future at the intersection of AI and human intelligence is filled with promise and challenges. The synergy between human creativity, ethical judgment, and AI's computational power opens the door to unprecedented innovations, potentially revolutionizing industries and addressing complex global issues. Yet, these advancements also prompt us to reflect on their impact on society, from the evolution of the job market to new ethical considerations. As we move forward, our focus should not only be on technological progress but also on understanding and preparing for these changes. This journey represents more than just the advancement of technology; it's about forging a future where AI amplifies human capabilities, complementing rather than overshadowing the unique qualities of human intelligence.

Authors:

Aldo Ramírez , Ph.D., Responsible AI Specialist and Sr. Data Scientist at Wizeline

Aníbal Abarca , Chief Technology and AI Officer at Wizeline

This article was developed with the assistance of ChatGPT, an AI language model created by OpenAI. ChatGPT provided support in generating ideas, structuring content, and refining the language used in this piece.

Image prompt: "A gritty, edgy Brooklyn building wall covered in intricate graffiti art, showcasing abstract representations of music, art, computing, health, knowledge, and artificial intelligence. The graffiti is designed with a raw, urban style, reflecting the dynamic influence of AI in a more rugged and street-art aesthetic."