Many companies are blindly following the bandwagon of including AI extensively in their operations to cut costs without adequately focusing on the safeguards to manage it effectively. However, adopting AI without strict control is a recipe for disaster that can damage their reputation and may cause severe financial loss. These agents are trained to handle inquiries, initiate payments, and even manage investments of their clients much faster than any human can do.
All this looks good, and the company is making a good profit by cutting its operational costs. But a new scenario emerges wherein a sophisticated hacker manipulates instructions to this AI agent, and as it lacks human oversight, it can be manipulated to bypass security checks and execute transactions that are unauthorised. By the time people in authority realise what happened, the financial and reputational damage to your company is already done.
This is not some fantasy or a dream scenario, but something that the UK’s Financial Conduct Authority (FCA) has flagged for the global financial sector right now. The FCA has asked institutions to share their feedback about serious AI failures and also near misses wherein an AI agent on its own initiates something that can cause systemic disasters.
The ANZ Rush to Automate
At Cybernetics GI, we see that some major banks across Australia and New Zealand are going at full steam to implement these systems. Their main focus seems to be on reducing their human workforce and cutting operational costs. To do this, they are increasingly giving complex decision-making in their day-to-day operations to artificial intelligence, without providing stringent safeguards. Here we see a threat looming on the horizon: when you introduce autonomous technology directly into financial data streams, it creates new vulnerabilities that are currently not fully understood by specialists.
We suggest that our client go through rigorous cyber security testing to find out whether the algorithm they are trying to implement in their operations can be manipulated or not, and if it can be manipulated, what safeguards can be put in place to make it foolproof.
As experienced cyber security auditors, we carefully find control gaps in the workings of the AI and report them to the executive board so that they can train their human workforce to spot these cyber threats that the machines can sometimes overlook.
Regulatory Patience is Running Out
The regulators are not giving financial institutions or banks any leeway in case the AI they have adopted for their operations makes any mistake, due to which a financial loss or data breach occurs.
Australia: APRA and ASIC
The Australian Securities and Investments Commission (ASIC) and the Australian Prudential Regulation Authority (APRA) are closely focusing on AI governance. The regulators, alongside certified essential eight security auditors, expect financial institutions to have strict oversight. If any bank deploys an AI agent then it falls on the part of the board duty to specify which executive will take the blame if the AI for some reason breaches any data privacy laws.
New Zealand: FMA and RBNZ
Just like Australia, in New Zealand, the Financial Markets Authority (FMA) requires all existing fair conduct and Anti-Money Laundering (AML) obligations to apply directly to automated systems. The Reserve Bank of New Zealand (RBNZ) warns that if a bank or financial institution relies completely on third-party AI providers, then it creates a massive systemic risk for them.
Objective Visibility for the Board
Experts have said that when you replace staff with AI, it can save some money in the short term; however, a single autonomous error on its part can completely erase those savings. Therefore, the boards need to be very objective when figuring out whether complete handover of operations to AI without human supervision is a risk worth taking.
Financial institutions and banks must therefore focus on in-depth vulnerability assessment protocols before they let these AI tools go live in their daily operations. To have greater robustness and clarity in your thought process, we suggest a formal independent cyberse curity audit provided by a professional cybersecurity expert like Cybernetic Global Intelligence. Our experts carefully analyze your digital environments and the overall training of your staff to give you a complete picture of the security gaps in your system. So that you can address them first before your regulators or threat actors find them.