Fact or fiction? Separating the AI fantasy from reality

My goal was to make my AI character feel as threatening as possible. The pace of my novel relied on that – on how high I could make the stakes feel to readers. On how believable I could make what Silicon Valley folks often joke about as P(doom), i.e. the numerical chance that AI may lead to a catastrophic event.

Flash back to the present, to reality and I've been working on a presentation for clients. An advertising idea for a big brand, created with AI tools as part of the creative process – and we will probably use more AI tools for the production, too. Likely, media companies will use AI to distribute it, and then we will use AI to analyze the data as the campaign runs.

I look around, and there is more. Next to me, someone is on a Zoom call, with the system transparently filtering the noise of the room: that’s AI. On the other side of the room, two other people discuss how bad traffic was over the weekend – their GPS systems used AI to help them find the optimal route home.

From Terminator to The Matrix and everything Asimov, AI has captured the imagination of writers, filmmakers, and scientists alike. It has led to numerous fictional portrayals of its capabilities and intentions – causing widespread fear and the inability to differentiate between fantasy and fiction when it comes to AI. While these works often provide thrilling narratives, now that AI is becoming such a massive part of our daily lives, it is essential to distinguish between fictional depictions and the realities of AI.

At its core, there are five key differences between AI in fiction and its real-world counterparts:

The sentient monster

Fiction: Many fictional narratives depict AI as conscious beings with emotions and self-awareness, capable of experiencing the world on par with humans.

Reality: Present-day AI systems lack consciousness and sentience. They operate based on learning algorithms and data processing, devoid of subjective experiences or true understanding. AI systems excel at pattern recognition and complex computations, but all the meaning coming from their creations still comes from human interventions and decisions.

General intelligence v narrow AI

Fiction: Countless stories portray AI as possessing general intelligence, capable of performing any intellectual task with ease, and surpassing human abilities in every domain.

Reality: Current AI systems primarily exhibit narrow or specialized intelligence. They excel at specific tasks, such as image recognition, natural language processing, or playing games, but they struggle to generalize their knowledge to different contexts. Achieving true general intelligence remains an ongoing challenge. As of now, AI is less of an evil genius and more like a very confident, and slightly hungover intern.

Machines that feel

Fiction: In many fictional works, AI is portrayed as having emotions and empathetic capabilities, able to form deep emotional bonds with humans… or even hate us too.

Reality: While AI can simulate emotions to some extent through sentiment analysis or voice modulation, it lacks genuine emotional understanding. AI's interpretation of emotions is based on algorithms, making it fundamentally different from human emotions driven by complex cognitive and physiological processes. That’s not to say that the work it produces can’t feel emotional. It can. Anyone who has played with it knows it. But that emotion is additive. It comes as an extra layer, either from elements added by humans or by emotions we add to it as an audience.

The overlords

Fiction: Popular culture often portrays AI as autonomous entities capable of defying human control, leading to scenarios of rebellion or domination.

Reality: AI systems in reality operate within strict boundaries set by their programmers. They cannot autonomously override their directives or acquire self-awareness. The difference between old-style computing and now is that the control we have is no longer on the steps it will take and the skills we will teach but on the goals it has and the machine’s ability to learn to perform better. Learning, rewards, and goals are the keywords here. That’s the core of this new style of computing and where our ability to control it lives. Yes, these computers will develop skills we never programmed them to have, and they will surprise us often, but their action will always remain tied to the goals we programmed them to perform, motivated by the rewards we give them for getting closer. Any threat coming out of their skills, I’m sorry to say, will have to come from the initiative of us humans.

New moralities

Fiction: Many fictional narratives delve into ethical dilemmas surrounding AI, exploring questions of morality, human rights, and the relationship between AI and society.

Reality: While ethical considerations are vital in the development and deployment of AI, they are often more complex and nuanced than portrayed in fiction. Real-world AI faces challenges such as bias in data, privacy concerns, and transparency in decision-making algorithms. From a technological development standpoint, these issues require careful consideration and responsible design to mitigate potential negative impacts. From a societal perspective, it will require us to revisit some core ideas and perceptions of how we operate as a society and what defines us human beings. Ideas like intellectual property will be challenged and questioned in multiple ways. The difference between learning and copying will need to be reassessed. The very concept of creativity, art, and emotions will need to be reconsidered too in this new context. But that’s what’s expected from any significant technological leap anyway, right?

The technological wave of AI is coming and is unavoidable. There will be issues, but they won’t be those we are used to seeing in sci-fi books and superhero movies. They will be more accidental evolutions, specialized routines that take their goals too far before we realize the threat.

To prevent those, we must clarify how distinct AI’s fictional and real sides are. To grasp the imagination of the audience, stories need to bring a thrill. They need to bring images and make stories look spectacular.

But for those trying to understand how this new technological wave will impact the world and their work, the pyrotechnical debate or reading exciting reports and newsletters won’t help much.

One thing all practitioners seem to agree on is that AI is a very experiential kind of technology.

One you need to use, to try, to have a glimpse of the real future ahead. So the next time you see a new “specialist” trying to explain the future based on things they read, ask them what are the most advanced experiments they have made themselves. Not the attempts they asked their intern to do, the ones they physically participated in themselves.

Then ask how that experiment changed the way they see the future. Because that—real experience, is the only reliable way to distinguish the most delicious popcorn fiction from the most actionable understanding of our upcoming reality.

Instead of doom, I see leaps. Climate, diseases, economy… there's a lot to gain, and the pitfalls — the democracy issues, the copyrights discussions, the wealth concentration, and bias concerns — I am sure we will be able to manage.

In the ad business, there are a lot of opportunities too. Advancements in ethical targeting. New forms the creative process can take. New ways we’ll have to gather insights. In the real world, everything is about to change, and my experiments so far make me very optimistic and excited about it.

What do yours say?