The optimism of #Purpose & science of #design in forging #digital customer experiences! Opinions my own!

By Luke Soon

In the ever-evolving realm of artificial intelligence (AI), the boundaries between reality and speculation are continuously shifting. While most of today’s breakthroughs remain confined to narrow AI—systems optimised for specific tasks—we are inching closer to an era where more generalised forms of AI are becoming conceivable. Among the theoretical constructs often discussed in both academic and practical circles are Theory-of-Mind AI and Self-Awareness AI. These concepts, though hypothetical, serve as powerful thought experiments that challenge how we design, deploy, and coexist with intelligent systems.

What is Theory-of-Mind AI?

Theory-of-Mind AI is inspired by the psychological concept of “Theory of Mind,” which refers to the ability to attribute mental states—such as beliefs, desires, and intentions—to oneself and others. In AI, this would mean developing machines capable of understanding human emotions, thoughts, and behaviours on a deep, contextual level.

While strides have been made towards this goal, such as MIT’s Kismet robot or Hanson Robotics’ Sophia, these attempts are rudimentary at best. For example:

• Kismet, a robot head developed at MIT, can recognise and mimic human emotions but lacks the ability to follow a human gaze or pay sustained attention.

• Sophia, a more advanced humanoid, can make eye contact, recognise people, and engage in basic conversational exchanges. Yet, it still falls short of truly understanding the human condition, much less predicting nuanced human reactions.

Theory-of-Mind AI, when fully realised, would allow machines to engage with humans in profoundly empathetic and personalised ways. Imagine an AI that not only detects stress in your voice but also understands the underlying causes—be it professional burnout, personal loss, or an unspoken fear—and responds appropriately.

The Allure and Challenges of Self-Awareness AI

Self-Awareness AI pushes the envelope even further. This concept envisions machines that possess a form of consciousness, including their own emotions, needs, and beliefs. Matthews (2017) provides a vivid illustration of how such an AI might behave:

Picture a self-aware robot shopping for groceries on behalf of its owner. Not only does it remember the owner’s preferences, but it also feels excitement over a sale on a favourite product, disappointment when stock is unavailable, and even impatience when faced with long checkout lines.

The idea of self-awareness in machines is compelling, yet it raises critical questions:

• What is consciousness in the context of AI? Can emotions or beliefs in a machine ever be truly analogous to human experiences?

• Ethical Implications: Would creating a self-aware AI imply granting it rights or protections akin to those of humans or sentient beings?

• Control and Oversight: If such an AI surpasses human intelligence and develops its own goals, how do we ensure alignment with human values?

Current State and Roadblocks

Neither Theory-of-Mind AI nor Self-Awareness AI exists today, and the journey towards their realisation is fraught with technical, philosophical, and ethical challenges. For instance:

1. Complexity of Human Emotions and Cognition: Understanding the full spectrum of human thoughts and behaviours requires grappling with variables that even humans struggle to comprehend.

2. Ethical Boundaries: The very act of creating systems that mimic or replicate human consciousness demands a rigorous examination of moral boundaries.

3. Technological Limitations: Existing models, while powerful, are still grounded in narrow applications and lack the generalised reasoning capabilities that these hypothetical AIs require.

Why These Theories Matter

Despite their hypothetical status, Theory-of-Mind AI and Self-Awareness AI offer valuable frameworks for thinking about the future of AI. They compel us to ask not only what AI can do for us but also what kind of relationship we wish to have with it.

As an AI ethicist and advocate for Human Experience (HX), I believe these discussions are essential for shaping a future where AI augments rather than overshadows humanity. If we aim to design systems that resonate with our human values—empathy, trust, and responsibility—then Theory-of-Mind AI serves as a guiding light. Meanwhile, the concept of Self-Awareness AI forces us to confront deeper questions about our own humanity and the ethical frameworks that underpin it.

Looking Ahead

While planning for Artificial General Intelligence (AGI) and beyond may feel like science fiction, the journey starts with the questions we ask today. Whether developing algorithms that approximate empathy or contemplating the ethical frameworks for AI self-awareness, these thought experiments push us to think boldly and responsibly.

As we continue to explore this brave new world, the ultimate challenge will not merely be technical but existential: how do we create intelligent systems that reflect the best of humanity while safeguarding against the worst?

In the words of OpenAI: “The future doesn’t just happen—it’s built.” Let us build it wisely.

Luke Soon

AI Ethicist, Partner at PwC, and Author of “Genesis: Human Experience in the Age of AI”