Subscribe to our FREE Newsletter, or Telegram and WhatsApp channels for the latest stories and updates.
In his book Talking To Strangers, Malcolm Gladwell writes about how a Harvard economist, 3 scientists, and a bail expert created an AI tool in 2017 to make decisions if accused individuals can be allowed to be released on bail.
The AI was given 554,689 cases, the same information the courts had about the accused, and was commanded to make a list of 400,000 people who should be allowed bail.
The point of this experiment was to see if a human judge or the AI tool can better determine if an accused is “high-risk” — meaning whether it is likely for the accused to commit another crime while on bail.
The people on the computer’s list were 25% less likely to commit a crime while awaiting trial than the 400,000 people released by the judges of New York City… In the bake off, machine destroyed man.
Malcolm Gladwell, Talking to Strangers.
In Mercy (2026), Detective Chris Raven (Chris Pratt) is strapped to a chair facing an AI judge, Judge Maddox (Rebecca Ferguson). He is on trial for murdering his wife, Nicole (Annabelle Wallis).
He has 90 minutes to prove his innocence or he dies.
It’s 2029. Chris was an avid supporter of the establishment of the Mercy Capital Court — a court that uses AI judges for factually accurate verdicts. Every bit of information — camera footage, emails, any kind of online record, whether private or public — is at your disposal if you are on trial.
The thing is, Chris has a drinking and a temper problem. He punched the doorbell cam off when Nicole refused to let him in the house. Moments later, when his daughter Britt (Kylie Rogers) comes home, we hear her bone-chilling scream coming from inside the house.
She’s just found her mother’s body in a pool of blood. Meanwhile, CCTV footage placed Chris in a bar, getting drunk.
The next thing he remembers is waking up in the Mercy seat.
Despite the mounting evidence against him: His footprints on the crime scene, her blood on his shirt, almost all the available footage, his old drinking habits and anger issues creeping back again — Chris is certain that he did not murder his wife.
But Judge Maddox is not.
In Gladwell’s book, he also writes about how we think that, as human beings, we can make better judicial decisions, just because a human judge has the advantage of seeing the accused standing before them and looking them in the eyes. The AI tool obviously cannot do that, and give a verdict purely based on the facts alone.
Mercy explores this dilemma of using the soulless and inhumane to make decisions about real people’s lives and futures.
It’s hard to say that it’s a sci-fi film when it’s barely fiction anymore. AI in judiciary systems isn’t exactly a very far-fetched concept, even in Malaysia.
Sabah and Sarawak have been using AI in court sentencing since 2020. They don’t replace human judges (yet 
), but are used as guidelines and to help spot sentencing trends.
Mercy uses the screenlife technique to show the video evidence and calls.
Screenlife is basically a screen-within-a-screen format. Movies like Unfriended (2014), Searching (2018), and Missing (2023) are examples of screenlife films.
There were several points in the film when Judge Maddox pulled up videos of Chris that made me go: “Why would anyone be filming this?”
But then I looked through the 1467 random screenshots and shoddy videos that literally have no purpose being in my camera roll, and realised I could ask myself the same thing.
It isn’t just our relationship with technology that the film prods at, it’s also our relationship with each other.
Chris’ relationship with the people closest to him: his Alcoholics Anonymous (AA) sponsor, his daughter, and his police partner, are all tested in the 90 minutes.
How do we respond when someone we know is suddenly charged for murder? What happens when the person we thought we could rely on ends up betraying our trust? The blurring lines between human and technology bleed into their relationships with each other.
If you watch Mercy looking for definitive answers to the AI vs mankind dilemma, you won’t find any.
There are opinions claiming that this movie is pro-AI, and I think it’s because the movie does seem to level the playing field in that humans and AI both “make mistakes” (Chris literally says this in the final scenes).
But I didn’t see Mercy as pro-AI. If anything, it’s a movie that doesn’t really try to take a stand for anything. It’s pro-thinking-for-yourself versus being at the mercy of someone, or something, else.
The major issue with this movie is its final 20 minutes.
Chris’ closing lines about how “AI and human beings both make mistakes and we need to learn from them” is not entirely wrong (AI learns based on what we feed it), but it’s framed in a way that can feel nonsensical.
When Judge Maddox slowly comes to terms with Chris’ innocence, as he uses his very human, detective prowess to defend himself, she tries to help Chris and we see glimpses of her humanising. Which, again, isn’t far-fetched.
AI models these days are programmed to feel like companions. People literally have conversations with ChatGPT like they’re friends.
These cinematic depictions don’t necessarily equal pro-AI, it may very well just be an attempt to reflect an uncomfortable reality. The imperative is on us to decide if this is the future we want for ourselves, or simply fight to preserve the essence of our humanity.
At the end of the day, the experiment outlined in Gladwell’s book is just that — an experiment.
How can we know for sure if someone will commit a crime again? Who can properly determine the futures of the accused?
What we really didn’t need to see was that flying police bike. That was far-fetched. 
Mercy is currently playing in cinemas nationwide.
Share your thoughts with us via TRP’s Facebook, Twitter, Instagram, or Threads.
