Why an 'AI judge' should always remain in the realms of science fiction
No algorithm can comprehend trauma, ethical nuance or the potential for rehabilitation in those entering the justice system
I had a conversation with a former colleague who spoke passionately about how artificial intelligence has transformed his life. Living with dyslexia and navigating the endless stream of written reports in his professional world was a daily struggle for years.
Now with AI at his fingertips, that burden has lifted. His stress has eased, his wellbeing improved, and he feels more in control. For him, AI has been a great leveller.
I’ve always been an early adopter – fascinated by AI’s potential, especially in medicine. I enjoy listening to those working closely with these technologies describe their applications and possibilities.
Earlier this year, a version of AI passed the Turing Test, meaning it could convincingly mimic human conversation. It doesn’t mean the machine is sentient, but it does raise questions about where we’re headed.
AI’s lack of depth
Yet, there’s something about the polished perfection of AI-generated content that leaves me cold. Emails, CVs, and job applications are beginning to feel too perfect, too uniform. I find myself appreciating the misspelling, imperfect grammar or random capitals – signs a real person crafted the message.
A friend, a brilliant writer, reminded me that it takes 10,000 hours to master a skill. With AI, the appearance of expertise can be achieved in seconds. But it’s a veneer, often not backed by depth.
The World Economic Forum’s 2025 report on the future of jobs, shaped by AI, climate change, and economic shifts, included a sobering statistic: 59 per cent of the global workforce will need reskilling or upskilling by 2030.
While new roles will emerge (as always) – the pace of change is dizzying. The real challenge lies in how we, as humans, keep up. And what becomes of those who can’t?
In the justice sector, AI is already making its mark. It’s being used to transcribe interviews, support justice social work, and streamline some court processes. These efficiencies are welcome. The accuracy of these tools is remarkable, but human oversight is still required.
A deeply human endeavour
When AI is used to aid decision-making or support policing, the implications become more complex. I watched a series of promotional videos showcasing AI in justice. They were slick, persuasive, and full of promise: faster decisions, smarter systems, predictive policing. The message was clear: we can all “be better”.
But AI is only as good as the data it’s trained on. In justice, that data comes from decades of reports, research, and systems, many of which reflect historical biases, systemic flaws and human rights abuses.
Justice is a deeply human endeavour. It affects lives in profound ways. How someone speaks in court, their body language, the emotion behind their words – these are things we instinctively read and interpret as humans. Machines can’t replicate that.
No algorithm can comprehend trauma, ethical nuance or the potential for rehabilitation in those entering the justice system. The idea of an AI judge belongs, I hope, firmly in the realm of science fiction.
Human judgment must remain at the heart of justice. I know the genie is out of the bottle. Tech entrepreneurs are reshaping the world at breakneck speed. But I believe no dataset, no matter how vast, can capture the full complexity of human life – its courage, vulnerability, and capacity for love.
This column was written by me, Karyn McCluskey, chief executive of Community Justice Scotland