Well, someone finally did it. Viewers of Channel 4’s Dispatches: Will AI take my job? found all was not as it seemed as the programme drew to a close. The experiment pitted four professionals against AI, seeking to determine whether humans or machines would prevail in a series of work-based challenges. Rather unsettlingly, the digital creations were represented by convincing replicas of the humans they were competing with. The four, a doctor, musician, lawyer and photographer, were all visibly shocked upon first encountering their avatars.
The documentary marked journalist Aisha Gaban’s screen debut. A straight-from-central-casting presenter, easy on the eye and articulating fluently in accentless English, Gaban handled things with aplomb as the plucky foursome battled their algorithmic adversaries.
Thankfully, for those of us with hearts that beat, the result went our way. Flawed though we undoubtedly are, it was the organics ‘wot won it’, narrowly outperforming the digital upstarts. It was, however, no pushover. AI did not disgrace itself, demonstrating how close it is to at least achieving parity, with all the potential repercussions that outcome might unleash.
Underlining the significance of what had just happened, Dispatches saved its big reveal to the end, when Gaban made a startling confession. Rather than being a hitherto unseen television anchor, it announced that it was itself an AI creation.
Channel 4 was careful to point out that the stunt was under control. By revealing the robot, it ensured adherence to the ethical policies it unveiled earlier this year, designed to ensure transparency in its use of the technology. Perhaps surprisingly for a broadcaster not known for its reticence to court controversy, it seemed to fear a public outcry.
It needn’t have worried. If the producers hoped for howls of fury, they were disappointed. Most of the comment around the event seemed to come from within the media and entertainment industry itself where, possibly unwittingly, Channel 4 had tapped into a rich seam of strife. Concerns around the use of AI—more specifically, generative AI—are growing. Barely a week passes without a doom-laden prediction of mass unemployment brought about by competition from the robots. From Croydon to California, workers in the creative industries are increasingly worried about the march of the machines.
But do they need to be? First of all, let’s not forget that AI isn’t really anything new. Sure, the mass collection and processing of data and with the power to process it is increasing at an exponential rate, but as Moore’s Law tells us, that was always an inevitability. What maybe wasn’t so inevitable was the technology’s ability to fool us—or, more accurately, our ability to fool ourselves.
Losing the imitation game
Arguments around the ability of machines to pass the so-called Turing Test, in which “an average interrogator will not have more than 70 per cent chance of making the right identification [human or machine] after five minutes of questioning” still rage, even after UC San Diego researchers achieved a 73 per cent rate earlier this year. Indeed, some experts suggest Turing’s ‘Imitation Game’ is no longer a reliable measure.
The technology behind AI is certainly ingenious, its capacity for data processing phenomenal (as is our own), but it just isn’t human. Aisha Gaban might look and sound the part, but it is simply incapable of reacting with empathy to an upsetting story. Nor can it hold a captivated audience in the palm of its hand as it relays the still-breaking details of a world-changing event, or leap excitedly out of its chair to cheer on a record-breaking athlete closing in on a gold medal victory. Gaban would never have “thought” to count them all out and count them all back again.
We have been here before. When the Luddites smashed the machinery of the Industrial Revolution, they did so because they feared their jobs would be taken. Instead, the world of work evolved. Humanity co-existed with its mechanised assistants and they jointly advanced our species.
While there are understandably concerns over AI, and regulation is urgently needed to head off a Wild West-style free-for-all nightmare, we would all do well to remember it is just another machine. AI is an incredibly powerful tool, but we built it. It exists to serve us. And, at least so far, it just can’t do what we do – it can’t successfully replicate spontaneous, emotionally-driven human behaviour.
Can it?
- This article originally appeared in the December issue of TVBEurope, available to download here.
