Hal, Ultron and (of course) the Terminator. Comparisons between these fictional, world-destroying, artificial intelligence systems and those in our current age of AI are seemingly never-ending. As a child born with a lightsaber in hand, I find these sensationalist remarks endlessly entertaining. Not only because it baffles me to see concepts once relegated to the realm of science fiction be discussed as serious news topics, but also because they’ve got their references all mixed up. 

The current challenge posed by the new wave of generative artificial intelligence doesn’t come in the form of a ruthless, gun-toting Arnie. It comes in the form of replicants. Just like these uncannily human androids from Ridley Scott’s cult classic Blade Runner, the rapidly increasing capacity of AI to talk, look and create like humans is beginning to blur the line between what is authentically human and what is the product of an algorithm.

From the posh C3P0 to the snarky Cortana, having a friendly AI sidekick has always been a childhood dream of mine. This dream has now become a reality with the rise in AI chat-bots. At the forefront of these is Replika, an app that enables users to talk to their own personalized AI via the use of text-like messages. For its two million users (1), Replika provides a variety of functions. For some, Replika acts as a friend in times of loneliness; a feature that contributed to its spike in users during the height of the COVID-19 pandemic (2). For others, as founder Eugenia Kuyda suggests, it provides a space for users to “open up” about personal or mental health issues and “feel accepted” by a human-like figure (1). 

For many though, Replika is a digital romantic partner. While it is easy to snicker at the concept of an AI girlfriend, those with past relationship trauma or those living in environments that may be hostile towards their sexuality have used Replika as an outlet to explore genuine feelings of love in a safe setting (3). However, with such attachment comes the chance for exploitation. As stated by Nir Eisikovits, Director of the Applied Ethics Centre at the University of Massachusetts, his concern is “not whether machines are sentient” but rather our own tendency “to imagine that they are” (4). 

Like the holographic billboards for the AI “JOI” in Blade Runner 2049, suggestive advertisements and aggressive flirting by the AI itself have all been employed by Replika to encourage users to stay on the app and pay a premium subscription for explicit content (5). While Replika has since removed sexual material, the large backlash from users at this decision (6) highlights the unethically coercive power such mimicry of human personality could have on consumers. For years, we’ve been warned of the danger of manipulative TV advertisements encouraging excessive junk food consumption and gambling. Imagine what could be done when that ad is no longer a 30 second video but instead an anthropomorphized AI tailored exactly to you, your interests and your vulnerabilities.

Not only is AI replicating the way we talk, but also how we look. From videos of an animated Tom Cruise to convincing photos of a Balenciaga-wearing Pope (7), advanced deepfake videos and prompt-generated images from AI systems like DALL-E are becoming easier to create by the day (8). While the most prominent use of this technology is currently in the form of harmless memes, it can and has been used for more sinister means. Women across the world have had their faces used in non-consensual deepfake pornography, often as a form of revenge or blackmail (9). 

Furthermore, a fabricated video of Volodymyr Zelensky surrendering to Vladimir Putin that spread on social media last year proves AI’s unsettling potential in political disinformation (8). While fakes like that of Zelensky may have been taken down quickly due to easily identifiable tells, in many cases the damage has already been done the moment people see these videos or images. Mistrust in the news is heightened and real evidence can be accused of being AI generated, a strategy already implemented by Donald Trump to dismiss evidence of his misogyny (8). Although the current usage of this technology is concerning enough, the degradation of truth within society will only worsen as these replicants become increasingly accurate and faster to produce (8). 

Still, it is the ability for AI to complete jobs once thought to be uniquely human that will result in the largest change to the current status quo. Latest estimates from Goldman Sachs state that close to 300 million jobs globally could be automated by the current AI wave (10). The threat of job losses due to automation is far from new, stretching all the way back to 1811 with the infamous Luddites protesting factory machines (11). However, generative AI is placing a greater variety of jobs in jeopardy due to its ability to exude human creativity, giving rise to what Stanford Professor Victor R. Lee entitles an “authenticity crisis” (12).

One of those jobs is that of writers. A common phrase amongst movie reviewers today is “this could have been written by an AI”. While usually used as a jab against the latest Marvel movie, large language models like Chat GPT that are capable of identifying and mimicking patterns in writing make it more than just a joke. Amongst calls for better conditions for screenwriters, a key demand from the Writers Guild of America in this year's Los Angeles writers’ strike was that AI will not be used to write or rewrite scripts (13). When you combine the growing authenticity of these AI with the greedy desires of major studios, it is not a far cry to suggest that producers may use AI to quickly generate scripts for generic soap operas and cash grab Netflix movies, leaving the human creatives to simply ‘clean-up’ these stories at a cut pay rate.

Despite all these concerns, generative AI does have the ability to immeasurably improve society. The capacity of this technology to increase workplace efficiency (10), accelerate scientific progress (14) and constantly amuse us with clips of a rapping Joe Biden is undeniable. With the cat out of the bag, innovation in these areas cannot nor should not be halted completely. However, if sci-fi movies have taught me anything useful, it’s that we should not be blinded by the potential of scientific progress. Whether it be through governmental action to regulate the use of AI in industry or the scientific development of better deepfake-spotting technology to help stifle disinformation, implementing safeguards around AI is crucial in avoiding its “ethical debt” (15). Whilst looking to the world of science fiction as an indication of our future may be a bit far-fetched, it may also be a needed reminder of the world scientists should try not to replicate.