All the banks and finance houses in Jamaica extol the virtues of this technology will enhance operations.

Many now scoff at the old brick-and-mortar model where the bank manager and staff know you and work to keep your money safe.

With AI, that is a practice of the past, a quaint way banking used to be done. There is no need to foster trust and relationships; just leave it to the machine to do the work.

There are increasing instances of cyber fraud at financial houses; firewalls are unable to hold up, criminals are able to find a way to your money, and banks are incapable of stopping them.

Ponder this: How many bank robberies have occurred in Jamaica over the last ten years? How many cyber breaches have occurred at financial institutions over the same period?

Banks and finance houses are ceding control and decision-making to cyber software, and now it will be AI.

Why did breaches at Mayberry, NCB and SSL go undetected? 

Anthropic’s Mythos is now the latest AI model system, they say, that will take banking to another level. However, it is able to hack complex systems undetected by humans.

This has caused alarm, and U.S. President Donald Trump has called in the country’s top banks, the Treasury Secretary Scott Bessant and the Federal Reserve Chairman Jerome Powell to look into dangers posed by Mythos.

Mythos could cause systemic financial catastrophe with its Anthropic Claude Code able to write programs that cannot be stopped.. Anthropic has already said Mythos will have “ offensive” and “ defensive” cyber capabilities.

Mythos is able to hack the most sophisticated banking systems, hospitals, government services and utilities. It can deliberately crash computers – think HAL in Stanley Kubrick’s masterpiece,  2001:A Space Odyssey. 

Anthropic said: “AI models have reached a new level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities.

“ The fallout for economies, public safety and national security could be severe.”

On the dangers posed here by Mythos and other such AI models, Fortune wrote: “ Anthropic’s new AI  model Mythos is causing a stir among cybersecurity experts and policymakers. The company says its new model is so skilled at finding and exploiting software vulnerabilities that it’s too dangerous to be released.

“Most agree that AI-driven cyber capabilities have reached a dangerous tipping point. Even without the powerful new model, they say existing, publicly available AI models can already carry out sophisticated cyberattacks- sometimes in minutes. Researchers are concerned about both the scale and accessibility of AI-enabled attacks. Tasks that once required advanced expertise, like scanning code for vulnerabilities, testing, and preventing attacks, are increasingly being automated or semi-automated by AI systems. Attackers, even those lacking a high level of technical skills, can  now launch highly automated attacks across thousands of systems at once in a massive coordinated assault.”

Will banking lose the human element? Is trust even necessary? There is now a higher probability that you can get up out of bed, and your money will not be in the bank because the institution was bamboozled by AI technology.

Banking is built on safety and security, now that is no longer guaranteed with AI being able to operate unrestrained. 

 Jouk Pleiter, CEO of AI banking platform, Backbone, too has concerns about the application of AI in banking, more so how it interfaces with the customer.

“Every bank is racing to deploy agentic artificial intelligence (AI). The ambition is real and the technology is ready. But the pilots aren’t landing, and the reason isn’t the AI. 

“AI needs three things to operate inside a bank: a complete customer context, a shared source of truth and governed authority to act. On a fragmented foundation, it gets none of those. An agent can’t run on partial data, following rules from one system while another enforces different rules, writing back to a third that doesn’t know what the first two decided.

“Fragmentation was always expensive. In the AI era, it becomes destabilising. AI needs a single current view of the customer-what they hold, what they’ve done, what signals they’re sending. Spread that data across six systems and the AI works with a fragment of the picture, making decisions on incomplete information,” said Pleiter.