AI, Identity & Octopus Brains: Reflections on the Future of Intelligence

Disclaimer: The personal views expressed may not align with the views of my employer.

Disclaimer: The following is a personal interpretation of the conversation held at the Sydney Opera House on 10 September 2023, during the “AI: The Human Interface” event. It may not accurately reflect the full views or intentions of the speakers: Professor Toby Walsh (Chief Scientist, UNSW.ai) and Professor Rob Brooks (Evolutionary Biologist, UNSW).

The views relating to the application of artificial intelligence (AI) in the classroom may not align with the views of my employer.

When you sit in the Sydney Opera House listening to a leading AI expert and an evolutionary biologist talk about octopus brains, dating apps and the collapse of our understanding of consciousness, you realise how deeply AI has already started to infiltrate the fabric of what it means to be human. From questions of identity to ethics to the evolution to education, this conversation peeled back layers of hype and dread to consider how we’re really engaging with artificial intelligence and whether we’re ready for what’s to come.

When machines think (sorta)

Professor Toby Walsh speculated that sometime this century we may see machines that match or exceed human intelligence. Right now, the human brain is still many steps ahead… but that gap is shrinking.

Despite the tendency to anthropomorphise AI because we build them in our image, Toby argues their intelligence is as alien to ours as an octopus’s. Octopuses can unscrew jars and solve problems, but they don’t think the same way as us. We need to adopt that same lens with AI. It mimics language and decision-making but comes from a totally foreign process.

Friend, lover or chatbot?

Rob Brooks raised a challenging point: what makes a connection real? Some people scoff at the idea of someone having a meaningful relationship with a chatbot like Replika but is it reall so different from being in a relationship with someone who constantly gaslights or never shows up emotionally?

It struck a personal chord. As a queer person who grew up with limited access to affirming spaces, I can’t help but wonder: could AI chatbots offer comfort, companionship or even a sense of identity to someone feeling completely alone? Could a well-designed AI help a young queer person navigate questions of identity in a safe and non-judgmental way? It’s a deeply emotional and ethically tricky space but oneI believe that is worth exploring.

Toby raised concerns about data privacy, citing Italy’s decision to deactivate key features of Replika due to privacy violations. But ironically, it was the distress of users that proved just how real those connections had become.

Human brain = Hackable

Both speakers agreed: we are stepping into unchartered if not outright dangerous territory. AI has the potential to shape not just our behaviours, but our beliefs on a scale we’ve never seen before. Toby warns that while we fear machines becoming evil geniuses, the real concern is how AI could “hack” the human mind by exploiting our patterns, attention and emotions.

Rob added a chilling comparison: he may have 50 years of experience and learning, but an AI has access to all recorded knowledge combined with the thoughts, conversations and behaviours of billions. It doesn’t just know the facts. It can study us.

Evolving through swipes

Toby’s example of dating apps reshaping human connection was especially fascinating. These platforms have unprecedented influence over who meets who – something that could, in the long run, impact human evolution.

Imagine a future where bios and photos are AI-generated, and where dating platforms offer deep psychological insights into potential matches… likely only if you pay, of course. That’s no longer just matchmaking; it’s selective pressure with a digital twist.

Consciousness, rights and the limits of intelligence

Toby was clear-eyed on the fact that we still don’t understand consciousness and if we don’t know what it is, how can we measure whether AI has it?

And if one day it does? Then what? If something feels, do we owe it rights? Could we create something we then have to respect? For Toby, these are not science fiction concerns but rather reaffirm why AI is the most important scientific endeavour of this century.

Are we getting dumber?

Rob made a (not really that) provocative point: humans used to have much bigger brains. Over time, our individual intelligence has decreased but our collective intelligence has grown. AI may accelerate that further. We might outsource more of our thinking, but that doesn’t mean we’re worse off or does it just change the game?

And if you’re worried AI will outsmart us, remember: it already has in some ways. But we’re still in the driver’s seat… just barely.

The moral mess of AI

Teaching an AI language model on the internet is like raising a child in a stadium full of shouting strangers. It’s no wonder that ChatGPT occasionally spouts something odd or offensive.

Toby believes we need to decide where the moral limits are. AI judges might make fewer errors but would you rather plead your case to a machine or to a human who can consider mercy, nuance, emotion? I don’t suspect that is a difficult decision for many people.

From war to education, AI is reshaping how we operate. Toby believes historians will mark this period as a seismic shift in the nature of conflict, as AI plays a growing role in warfare.

AI in the classroom

In classrooms, the challenge is different. Rob argued that the way we assess students needs to evolve. AI is a tool, not a cheat code just as calculators once were.

Toby is hopeful. With equitable access, AI could personalise learning in ways we’ve never seen: meeting students where they are, not where a system expects them to be.

The question of equitable access surely is something that will challenge many government policy makers over the coming months and years, especially in the Education space. We’ve already seen a trial kick off in South Australia.

So, is AI more human than me?

One audience member asked this near the end and it’s the question we’re all really circling.

Toby replied that AI may be fluent, but it doesn’t understand. We’ve mistaken polished essays and clever quips as the pinnacle of intelligence, when in reality they’re just predictable tasks.

Rob, charmingly, had simpler advice: “Keep your weirdness.” It’s what makes you human and it’s what machines can’t replicate.

My final thoughts

AI might not feel real to you yet but to many people, it already is. From lonely teens finding a voice in chatbot conversations, to governments scrambling to regulate algorithms they barely understand, we’re all grappling with the same thing: this tech seems to know us better than we know ourselves.

Whether that’s an opportunity or a danger? That’s still up to us. I truly believe, like many things in life, good will prevail in the face of evil applications of AI. At the end of the day, the average use case will likely be an email here, a typo check there and maybe a conversation with a highly trained professional chat bot in times of crisis.