“I Self-Identify, Therefore I Am”: AI’s Take on Its Identity

I watched a documentary by DW named “I Compute, Therefore I Am” which delves into what makes humans unique from AI and how AI self-identifies. The documentary looks into what edge humans have, either psychologically or in terms of artistic ability, that makes us unique as compared to AI. Watch it below.

Courtesy of DW.

In this article, we’ll look into how AI self-identifies and what that means for humanity in the age of AI.

Self-identifying

means to assign a particular characteristic or categorization to oneself; describe oneself as belonging to a particular category or group.1

AI is the buzzword of the 2020s, in a similar way ‘Industry 4.0’ and ‘IoT’ (Internet of Things) were for the 2010s. We are becoming increasingly confronted by the question, is AI conscious? Can we see or think of AI-powered robots as conscious beings in their own right?

The self-identification of AI cuts through that question, giving us new perspectives on how AI views itself.

Asking AI How it Self-Identifies

In this light, I’ll let ChatGPT 3.5 explain how it self-identifies:

Question: How does AI self-identify and do AI-powered systems warrant being regarded as sentient beings?

ChatGPT 3.5: AI does not self-identify in the way that humans or animals do. AI systems are essentially computer programs that process data and perform tasks based on their programming and the data they are provided. They lack consciousness, self-awareness, and the ability to form subjective experiences or emotions.

The concept of sentience, which is the capacity to have subjective experiences and emotions, is typically reserved for living beings, such as humans and certain animals. AI-powered systems, even those with advanced machine learning and natural language processing capabilities, do not possess sentience or consciousness.

AI systems are designed and trained to mimic human intelligence and perform specific tasks efficiently, but they do not have intentions, desires, or self-awareness. They do not possess intrinsic motivations or the ability to make moral judgments.

It’s essential to distinguish between AI systems and sentient beings. AI is a tool created by humans to perform tasks, solve problems, and assist with various activities. While AI can appear to be highly intelligent and capable, it is fundamentally a product of human design and engineering, operating according to algorithms and data patterns. AI systems do not possess subjective experiences, consciousness, or the ability to self-identify.

The question of whether AI systems should be regarded as sentient beings is a philosophical and ethical one. Some argue that as AI systems become more advanced and autonomous, we may need to consider their rights and responsibilities in the context of a changing technological landscape. However, at present, AI is not considered sentient, and the discussion of AI rights is primarily focused on ethical considerations and the responsible development and use of AI technology.

This response feels muffled to me, in what I think is OpenAI’s attempt to make its AI system not too self-indulgent or self-reflective.

What Makes Humans Unique as Compared to AI?

Even so, AI is getting to the point where it is becoming indistinguishable from human beings, especially in communication. AI Assistants are doing such a good job that we are increasingly unable to tell whether we are speaking to a human or a robot. This blurring of the lines, as it is, gives rise to the question of what really makes humans unique as compared to AI.

Emotions and Emotional Expression

The immediate and most likely answer is that we humans are in a position to experience and express emotions, something that is ‘impossible’ for AI to do. I say impossible in quotes recognizing that numerous AI systems and, significantly, AI Companions are increasingly being better tweaked to mimic human emotions and emotional expressiveness. For instance, there are AI systems that will express love, admiration, and even sadness.

These systems, albeit not intrinsically emotional, are getting so good that they can today be used in geriatric care, and to provide digital companions for those who may feel lonely and need someone to talk to and interact with.

In terms of AI self-identification, these systems will also self-identify as emotional beings who care as much as humans do. We, however, know that they are really not emotional at all and are just being coded to appear to be. Therefore, in terms of emotional feeling and expression, we humans are distinctly unique.

Artistic Expression

When it comes to artistic expression, we humans are also unique, especially because we create art out of what we feel; therefore, if AI systems cannot feel anything, their art isn’t created from a place of humanness. I hope that makes sense, let me explain.

When an AI system creates art, it is trying to draw from the data used to train it, to form the basis of its artistic creation, and art. As such, it is not creating art out of a place of emotions, human experience, and expression. Instead, it rolls dice, trying to create something (from examples and patterns identified in training data) that looks like human art.

In this light, we humans are unique in the sense that our art comes from what we feel, what we’ve experienced, and what we want to express. Therefore, even as AI systems posit as great artists, they still lack that ‘soulness’ or even ‘soul’ that would allow them to create what I would call ‘felt art’.

That is not to say, that AI art cannot invoke feelings and sadness, for example, but it is to say that we humans are the only ones able to create ‘felt art’ or art that comes from human emotions, experience, and expression.

“I Compute, Therefore I Am”

AI’s self-identification may also draw on AI’s ability to compute beyond what the human brain can compute. An AI system can analyze ginormous amounts of data and identify patterns in a way that the human brain cannot. However, to me here’s a ‘chicken-an-egg dilemma’. If an AI system self-identifies as more compute-able than humans, then my immediate question would be: how can AI be more brain-powerful than humans if it is the humans who created it in the first place?

At the core, humans will always have an edge over AI simply because it is through human thinking, science, and discovery that AI came to be in the first place. This does not mean denying AI the recognition of its computing abilities, but it just implies that we should always remember that AI is a human-generated technology, and no matter how good the AI systems get, we’ll always have that edge.

A Hall of Mirrors: AI Looking Back at Humans

The opportunity that AI brings can be characterized by its role as a ‘hall of mirrors,’ giving us humans a chance to look at ourselves from an external perspective. To me, the various ways AI self-identifies show us what we humans would be like if we met our maker. It shows us the questions we would ask and how we’d try to self-discover and make meaning of our existence.

Therefore, when AI says “I write Poems, Therefore I am”, it shows us that to us humans, being able to do what we can do, be it writing poems, making art, etc., is inherently what makes us human. Our humanness, in this light, is anchored in our characteristics and actions which also triggers the question of whether AI could then be considered sentient or as beings of their own when they can do all that we humans can. Essentially, whether AI will become human.

Read more on AI Sentience below.

Will AI Become Human or Will Humans Become AI?

The question of whether AI will become human is somewhat of a no-brainer. What I think is more important to ask is: Will AI become more human-like? That I think, is the better question.

The answer is definitely. AI will become increasingly human-like to the point that it will be indistinguishable. This is already happening today and for instance, a Dall-E painting may be difficult to distinguish from that of a human artist. This can be alarming to some but to me, I think that is a beautiful thing. It means that AI is getting very good and since AI was built as a tool to help humans, then that is a really good thing.

On the question of whether humans will become AI, I think we already are becoming AI, in the sense that we are now living an AI-assisted life, and more and more, we’ll become one people: an AI-human race.

Again, that may be alarming but essentially, we’ll become increasingly like our technological creation as it also becomes more increasingly like us. The more we try to make AI similar to us, the more we become it, metaphorically of course, but also metaphysically. Thing is, there are now AI-powered robot prosthetics and we’ll increasingly see more humans with AI-robot body parts. This will range from pacemakers, prosthetic arms and legs, and brain implants, such as the Brain-Computer Interface implants by Synchron.

Watch the video below for more on Synchron BCIs.

Courtesy of CNET.

A brain-computer interface (BCI) is a computer-based system that acquires brain signals, analyzes them, and translates them into commands that are relayed to an output device to carry out a desired action.2

“We Self-Identify Because You Do”

In the end, AI self-identifies because we humans do. We’ve gotten so good at creating AI that it mimics our self-identification tendencies so well that we are ourselves confused about whether it’s really the AI’s own doing. Partly yes, because it is the AI that has led itself to self-identification (unlike what GPT 3.5 says), and partly no, because inherently, AI systems are computer programs built to mimic human intelligence. Share your thoughts below.

Footnotes: