We asked AI to evaluate itself against the skills of a human. What grade will it give itself?
With all the advances in artificial intelligence happening at hypersonic speeds, it seems that the chances of an AI mimicking the abilities of a human brain no longer seem as remote or impossible as they were before the advent of large language models (LLMs) like ChatGPT. But how would AI score itself by comparing the human brain to its synthetic neural network?
We asked ChatGPT’s AI to create a list of 10 skills that would mimic those of a human brain, and then do a little big analysis of who would win in those skills or competencies in order to mimic a human brain. Will the AI believe that it has already managed to fully mimic our brain? Will it score itself appropriately? How long before this becomes a reality?
After these reflections, we will initially list the 10 skills in mourning: memory, perception, attention, reasoning, emotion, language, learning, creativity, motor control and problem solving.
To start, AI tells us that “comparing artificial intelligence with human capabilities is a fascinating exercise that highlights the divergent capabilities of machines and organic intelligences.” Considering that a score of 1 shows that AI would be far behind human capabilities and where a 10 would mean that AI equals or surpasses human abilities, these are the results:
Read Also:50 years of the Volkswagen Golf: more than a car, an emotional bridge between generations
Memory: AI: 10, Humans: 7
Table of Contents
This is where we lose. AI excels at everything that concerns memory and does not miss the result. It can store large amounts of information indefinitely and retrieve it without any hindrance. In contrast, human memory is fallible and subject to decay, interference, and bias. To tip the scales, AI values that human memory is associative and helps make connections between different pieces of information, something that even high-level AIs are still trying to replicate. AI gives itself an A+ above us.
Perception: AI: 8, Humans: 10
AI has made incredible advances in perception, especially in processing visual and auditory information. With the advent of machine learning and neural networks, AI is now able to recognise images and sounds with great accuracy. However, human perception is multimodal and deeply integrated and internalised, allowing for nuanced interpretation based on context, which is why AI still has some difficulties. It has given itself a high score compared to humans.
Attention: AI: 6, Humans: 9
Surprised? ChatGPT rates itself as somewhat poor in this regard, as human attention is highly adaptable and can effortlessly switch between tasks, albeit with limited capacity. Even though AI can process many streams of information simultaneously, it lacks the intuitive prioritization and selective focus that is innate to humans.
Reasoning: AI: 7, Humans: 9
AI can reason within the boundaries of logic and formal systems with superhuman ability, particularly in structured environments. However, human reasoning encompasses creativity, intuition, and moral considerations. Anything involving abstract reasoning is a bit outside of AI’s neural controls, so here ChatGPT is awarded an interesting 7 out of 10.
Emotion: AI: 2, Humans: 10
Interestingly, it doesn’t give itself a 1, but it is realistic in terms of emotion being a fundamentally human trait, which we use in a myriad of aspects of our lives: decision-making, empathy, or social interaction. AI can simulate emotional responses and recognize human emotions to a certain extent, but it doesn’t experience emotions – at least not yet – so it can’t genuinely relate on an emotional level.
Language: AI: 7, Humans: 10
As expected, given that language is inherently human, we got an A and AI got a B. While AI has made progress in understanding and generating natural language through developments in NLP (Natural Language Processing), it often fails to capture the deeper nuances, cultural contexts and emotional subtleties that come so naturally to us.
Learning: AI: 8, Humans: 9
The result is very similar. AI can learn from large data sets at an incredible speed, which is a significant advantage over human learning . However, humans learn in a more versatile and less data-intensive way, says AI, which means that we can understand concepts from fewer examples and can apply them in a broader context.
Creativity: 4, Humans: 10
We are still more creative than AI. Creativity involves not only the generation of novel ideas, but also the ability to judge the value of these ideas and build on them in meaningful ways. While AI can generate new patterns and ideas (as in art or music), it lacks the intrinsic drive and subjective experience that fuels human creativity.
Problem Solving: AI: 8, Humans: 9
Is it curious that it hasn’t caught up with humans yet? AI is great at solving structured problems with clear parameters (e.g., debugging software). However, humans excel at solving dynamic, real-world problems that require context understanding, improvisation, and adaptability—capabilities that AI doesn’t possess.
Read Also:Underground car parks: the deadly trap of DANA?
Motor control: AI: 7, Humans: 9
Robotic systems have made progress in motor control, performing precise and repetitive tasks better than humans. But when it comes to the complex, adaptive and fine-tuned motor skills that humans possess (whether for activities like sports or art), AI and robotics have yet to match human agility.
Final AI grade: 6.7
This final average score on this 10-question exam reflects, according to AI, the strong performance of artificial intelligence in some areas (such as memory and learning) and its continued development and room for growth in others (such as emotion and creativity).
