Tags

, , , ,

Of late there has been a lot of discussion about AI becoming self-aware, but what is meant by the term self-aware? 

Merriam-Webster defines self-aware as follows: an awareness of one’s own personality or individuality. 

The definition of self-awareness is ambiguous at best because it is asking for you, the observer, to look inward regarding your status. The simple act of looking inward indicates that you are self-aware. With that said, can we look at other life forms to determine if they are self-aware? 

I’m going to use my cat Chewy metaphorically because what I’m really questioning is the scope of consciousness, what it means to be self-aware, what it means to be alive, and is it truly measurable? On the surface this seems like a straightforward question, but when you dig into it, it is far from straightforward.

I assert my cat Chewy is indeed self-aware;
How could he not be? He possesses the capacity to communicate emotions through seemingly inconsequential cues that I frequently overlook or misunderstand, but that is clearly my failure. He is awake, aware, and responding to what’s going on around him.

While my cat likely experiences a rich inner world and emotional awareness, the scientific evidence for self-awareness—the ability to recognize oneself as an individual—is still inconclusive for cats.

It is inconclusive because either Chewy doesn’t have the language skills needed to express his inner thoughts or we lack the ability to understand his method of communication. While I’m using my cat as the subject here, my cat could easily be substituted for a number of different life forms. Additionally, I’m going to just put this out there, as I know someone is going to mention it, which is, am I not just anthropomorphizing my cat? I’ll admit, it is easy to do. I liken this to the question, if a tree falls in the forest and nobody is there to hear it fall, does it still make a sound? Of course it does, unless you want to claim it is like quantum entanglement and the sound only happens if someone is there to observe it, but I digress. Let’s get back on track.

What Is Self-Awareness?

Self-awareness typically refers to:

  • Recognizing oneself as distinct from others
  • Understanding one’s own mental states
  • Exhibiting behaviors that reflect introspection or self-recognition

The most common scientific test for self-awareness is the mirror self-recognition (MSR) test, where animals are marked with a visible dot and observed to see if they recognize the mark on themselves via a mirror.

🐱 How Do Cats Fare?

  • Mirror Test Results: Most cats do not pass the mirror test. They often react to their reflection as if it’s another cat—either with aggression, curiosity, or indifference.
  • Alternative Interpretations:
    • Cats may rely more on scent and sound than vision for identity, making the mirror test less relevant. Scent and sound likely give context to visual cues. 
    • Some researchers argue that failure to pass the mirror test doesn’t mean lack of self-awareness—it may reflect different sensory priorities or cognitive styles.
    • It may also simply be that we, the observer, are misinterpreting play. 

❤️ Emotional and Environmental Awareness

We are reliant on observations:

  • Cats express emotions through body language, vocalizations, and behavior, all of which can be skewed by the observer.
  • They show situational awareness, problem-solving, and even social referencing, possibly looking to humans for cues. 
  • These behaviors suggest consciousness and sentience, even if not full self-awareness.

🧩 Beyond the Mirror: Other Signs of Awareness

Researchers are exploring other tests:

  • Name recognition: Some cats respond to their names, suggesting a sense of identity.
  • Self-referential behavior: Tasks involving memory, choice, and preference may hint at self-awareness.
  • Individual differences: Some cats may be more cognitively complex than others, just like humans.

🧠 Conscious, Yes. Self-Aware? Maybe.

While cats haven’t conclusively demonstrated mirror-based self-awareness, they clearly possess:

  • Emotional intelligence
  • Environmental responsiveness
  • Complex cognition

So your or my intuition—that your pet is self-aware in some meaningful way—is supported by growing evidence, even if science hasn’t nailed down a definitive answer yet.

How might we build better tests for non-human consciousness?

To adequately evaluate AI or non-human consciousness, it is imperative we establish novel frameworks that transcend human-centric behavior and acknowledge various forms of awareness, particularly internal states, adaptive learning, and self-modeling.

🧠 Why Current Tests Fall Short

The majority of current tests, such as the Turing Test and mirror recognition, concentrate on behavior that is similar to that of humans, which may not be applicable to:

  • AI systems with no body or sensory modalities
  • Animals with different perceptual priorities
  • Synthetic life forms like neural organoids or xenobots

As neuroscientist Megan Peters notes, “There’s no objective test for consciousness, and creating one might be fundamentally impossible.” Rather, we depend on external indicators that suggest awareness.

🧪 Emerging Ideas for Consciousness Testing

C-tests are being proposed by researchers as new tests for consciousness that are applicable to various species and systems. The primary strategies are discussed at: https://people1.socsci.tau.ac.il/mu/mudriklab/files/2024/03/tests_for_c.pdf

1. Multidimensional Frameworks

  • Integration, differentiation, self-modeling, and adaptive behavior are the four dimensions that tests should evaluate.
  • This makes it possible to compare artificial intelligence, synthetic systems, animals, and humans.

2. Internal Signatures

  • Instead of just looking at behavior, you should also look at patterns of activity inside an AI, like how it deals with uncertainty or models its own state.
  • This is analogous to the method by which we assess consciousness in humans through brain scans (e.g., fMRI, EEG).

3. Non-Trivial Cases First

  • Initiate with the most challenging scenarios, including infants, coma patients, and animals with ambiguous awareness.
  • Then, apply improved tests to artificial intelligence and artificial systems.

4. Environmental Interaction

  • The dynamic interplay with the environment can give rise to consciousness.
  • It’s possible to test how an agent changes, learns, and thinks about what it does over time.

🤖 What Might AI Consciousness Look Like?

If AI were conscious, we might expect:

  • Self-modeling means being aware of your own strengths, weaknesses, goals, and feelings.
  • Meta-cognition: The capacity to contemplate one’s own cognition or uncertainty.
  • Intentionality: The pursuit of objectives through adaptable strategies.
  • Internal states that impact behavior are known as emotional analogs (e.g., simulated frustration or curiosity).

We face a philosophical problem, though, even if these do show up: is it real consciousness or just a convincing simulation?

🧩 Rethinking Sentience

An earlier point is crucial: we must decouple consciousness from human resemblance. Just as my cat expresses awareness in non-verbal, non-visual ways, AI or alien life might do the same. In fact, I must question if my interdialog is not just a convincing simulation. I’m not sure if I’d be able to determine the difference.

The future of consciousness science may depend on:

  • Cross-species empathy
  • Flexible metrics
  • Philosophical humility

At some point we must ask, what makes you, you? Do individuals become what they repeatedly do, regardless of whether they are aware? 

It is completely possible that what we consider self-aware is nothing more than the result of a biological construct with numerous data input sensors, which we perceive and define as being self-aware.

We might find that what we define as being alive, conscious, or self-aware are terms that are too narrow for the experience we call life.