From Humanoid to Human: The Evolution of AI

Artificial intelligence (AI) has stirred the imagination for decades. Seen as the ultimate antagonist in Stanley Kubrick’s 2001: A Space Odyssey, enslaving mankind in The Matrix, or, more recently, as a deceptive humanoid in Ex‑Machina, AI has become a classic pop‑culture reference. The abundance of AI representations in modern pop culture evokes the complex relationship that humans share with machines. Current status? It’s complicated. Our VisualGPS proprietary consumer data reveals mixed feelings: 50% of Europeans see AI as empowering, and 50% see it as a threat.

For organisations, this relationship is far less complicated. Today, AI adoption is widespread after years of progress, and the benefits for organisations are significant. Considering that European spending on artificial intelligence is poised to jump over $50 billion in 2025¹, it's no surprise that the AI effect is equally exponential in visuals. According to our recent research, the European tech industry's appetite for AI content has grown 34 times in the last five years.
The interest is real. However, despite the rapid adoption and headlines in marketing materials, AI as a visual subject remains challenging. The invisible and enigmatic nature of AI makes it imperceptible for tech brands that follow similar visual tropes over the years.

Do androids dream of digitally generated data?
You have probably noticed AI news stories illustrated with clichéd humanoid representations of AI. You have probably seen the famous panel from the Vatican's Sistine Chapel that shows the hand of God reaching out to give life… to a robot. Anthropomorphising robots was the predominant way that brands depicted AI five years ago, and to give it a human face or body was a simple visual way to demystify it. However, even with human features, the visual representation of AI was still reproducing visual codes known from some of the science‑fiction oeuvres—robots in cold and inhuman environments.

In 2021, the visual language around AI is radically different. The humanoid image that was so popular in 2016 has now been replaced with images of digitally generated data flows, holographic brains and enlightened motherboards. From super‑robots to ambient tech, from anthropomorphic to abstract, AI is now seen as a ubiquitous data flow, working tirelessly in the background. If digitally generated visuals are predominant, the classic blue and black tones overrule more surprising colour palettes. Nearly 8 in 10 images are futuristic concepts using blue as the primary colour, evoking speed, connection and technical prowess.

However, between robots and blue data flows, the visual language around AI is still missing the essential ingredient: humans are, most often, absent. While our attention is focused on robots, speed, connection or the technology itself, the popular imagery is disconnected from AI's significant societal implications.

Human after all
Why is it important? Because the tech industry has the power and responsibility to shape AI technology in ways that could benefit all. Therefore, it is essential to be intentionally inclusive by design and show who is represented in relation to AI. In recent years, much has been written about algorithmic bias. Through the work of Joy Buolamwini and the Algorithmic Justice League, we learn that to create unbiased AI, Who Codes Matters². Consequently, unless the tech industry puts effort into decoding their unconscious bias, they will replicate bias in every technology they create.

While we know that the tech industry does not consciously perpetuate stereotypes, our latest analysis of visuals used by the European tech sector reveals that unconscious bias exists in their visual choices. For example, women have 33% less visibility in programming or IT jobs than men. People outside binary gender identification are heavily underrepresented. The majority of popular people‑oriented visuals centre white people. If we go back again to science‑fiction references, the paper “The Whiteness of AI”³ argues that the depictions of artificial intelligence humanoids in pop culture as mostly white can carry a number of consequences, including the erasure of ethnic, sexual, gender or social diversity. The representation of AI as white humanoids situates machines in a power hierarchy, notably above currently marginalised groups.⁴

So how will the AI visual language evolve in future? If AI promises to benefit everyone, the logical evolution will be a more grounded depiction, highlighting the collaboration between AI and humanity. Today AI is shown as a separated or even independent entity, which is far from reassuring for those who have developed a negative perception of AI, which happens to be half of European society. So, when visualising AI, it’s important to choose inclusive visuals which emphasize transparency and AI benefits for all humans. Algorithmic bias is a problem and the root of the problem is not technological, but social. If all technology brands are committed to dismantling unconscious bias, we must be reassured about our future relationship with machines.

[1] European Spending on Artificial Intelligence Will Reach $22 Billion in 2022, Supported by Strong Investments Across Banking and Manufacturing, Says IDC (International Data Corporation)
[2] Leading a Cultural Movement Towards Equitable and Accountable AI (Algorithmic Justice League)
[3] Cave, S., Dihal, K. The Whiteness of AI. Philos. Technol. 33, 685–703 (2020). (Philosophy and Technology)
[4] Ibid.