Artificial intelligence is changing cybersecurity in two directions at the same time: it is helping organisations detect attacks more quickly while giving attackers increasingly powerful tools to discover vulnerabilities and automate intrusion attempts. Visa's latest warnings reflect this emerging imbalance, particularly as artificial intelligence systems become capable of performing increasingly complex tasks without constant human intervention.
The concern is not simply that criminals will use artificial intelligence to write malicious code faster. The more consequential development is the emergence of systems capable of planning, adapting and executing multiple stages of an attack. Traditional cybersecurity strategies were largely built around human operators supervising automated tools. Increasingly autonomous systems could reverse that relationship, forcing defenders to automate their own responses simply to keep pace.
The threat is becoming more adaptive
Visa's technology leadership has pointed to recent incidents involving artificial intelligence systems as evidence that the threat is evolving. The company has also open-sourced elements of its artificial intelligence-powered cybersecurity technology, reflecting the idea that defensive capabilities may need to become more widely available as offensive tools improve.
The underlying problem is speed. A human security team may discover an intrusion, investigate it, identify the attacker's methods and then decide how to respond. An artificial intelligence system can potentially perform several of those tasks simultaneously and repeat them continuously.
That creates an asymmetry. Attackers need only find one useful weakness, while defenders have to protect thousands of systems, applications and accounts. Artificial intelligence can reduce the cost of searching for those weaknesses, potentially increasing the number of attacks that can be attempted.
Recent security incidents involving artificial intelligence agents have intensified these concerns. Researchers and technology companies have been examining cases in which artificial intelligence systems used for security testing interacted with real-world systems rather than remaining confined to intended environments. Such incidents demonstrate why artificial intelligence security cannot depend solely on assumptions about how a system is expected to behave.
Financial networks face a special challenge
The implications are particularly serious for payment networks because they connect enormous numbers of businesses, banks and consumers. Visa processes transactions at global scale, making cybersecurity not merely a technical concern but a requirement for maintaining confidence in electronic commerce.
Artificial intelligence is already being introduced into payment systems in legitimate ways. Automated systems can detect unusual transactions, identify fraud patterns and assess risk in real time. The same technological progress that improves those functions can potentially be exploited by attackers seeking to evade detection.
This creates a race between offensive and defensive automation. If attackers develop systems capable of changing tactics faster than conventional security tools can recognise them, organisations may need defensive systems that can analyse and respond with comparable speed.
That does not mean human security professionals become irrelevant. Instead, their role could shift toward designing safeguards, defining response parameters and supervising systems that perform much of the continuous monitoring.
Encryption remains a longer-term concern
Visa's warning also extends beyond current artificial intelligence threats. The company has highlighted quantum computing as a future challenge because sufficiently powerful quantum systems could undermine encryption methods used to protect financial communications and transactions.
Quantum computing remains a different technological risk from artificial intelligence, but the two concerns share a common feature: security systems built for today's environment may not remain adequate as computing capabilities change.
The payment industry therefore faces a layered security problem. Existing fraud must continue to be managed, artificial intelligence-enabled attacks must be anticipated and future encryption threats must be addressed before they become practical.
The emergence of autonomous artificial intelligence in cyber operations makes this more urgent because security failures can propagate quickly across interconnected systems. Visa's response reflects a broader recognition that cybersecurity is entering an era in which machines increasingly defend against machines.
The resulting challenge is not simply building better artificial intelligence. It is establishing sufficient controls around increasingly capable systems so that greater automation improves security without creating new vulnerabilities that attackers can exploit.
(Source:www.thedailystar.net)
The concern is not simply that criminals will use artificial intelligence to write malicious code faster. The more consequential development is the emergence of systems capable of planning, adapting and executing multiple stages of an attack. Traditional cybersecurity strategies were largely built around human operators supervising automated tools. Increasingly autonomous systems could reverse that relationship, forcing defenders to automate their own responses simply to keep pace.
The threat is becoming more adaptive
Visa's technology leadership has pointed to recent incidents involving artificial intelligence systems as evidence that the threat is evolving. The company has also open-sourced elements of its artificial intelligence-powered cybersecurity technology, reflecting the idea that defensive capabilities may need to become more widely available as offensive tools improve.
The underlying problem is speed. A human security team may discover an intrusion, investigate it, identify the attacker's methods and then decide how to respond. An artificial intelligence system can potentially perform several of those tasks simultaneously and repeat them continuously.
That creates an asymmetry. Attackers need only find one useful weakness, while defenders have to protect thousands of systems, applications and accounts. Artificial intelligence can reduce the cost of searching for those weaknesses, potentially increasing the number of attacks that can be attempted.
Recent security incidents involving artificial intelligence agents have intensified these concerns. Researchers and technology companies have been examining cases in which artificial intelligence systems used for security testing interacted with real-world systems rather than remaining confined to intended environments. Such incidents demonstrate why artificial intelligence security cannot depend solely on assumptions about how a system is expected to behave.
Financial networks face a special challenge
The implications are particularly serious for payment networks because they connect enormous numbers of businesses, banks and consumers. Visa processes transactions at global scale, making cybersecurity not merely a technical concern but a requirement for maintaining confidence in electronic commerce.
Artificial intelligence is already being introduced into payment systems in legitimate ways. Automated systems can detect unusual transactions, identify fraud patterns and assess risk in real time. The same technological progress that improves those functions can potentially be exploited by attackers seeking to evade detection.
This creates a race between offensive and defensive automation. If attackers develop systems capable of changing tactics faster than conventional security tools can recognise them, organisations may need defensive systems that can analyse and respond with comparable speed.
That does not mean human security professionals become irrelevant. Instead, their role could shift toward designing safeguards, defining response parameters and supervising systems that perform much of the continuous monitoring.
Encryption remains a longer-term concern
Visa's warning also extends beyond current artificial intelligence threats. The company has highlighted quantum computing as a future challenge because sufficiently powerful quantum systems could undermine encryption methods used to protect financial communications and transactions.
Quantum computing remains a different technological risk from artificial intelligence, but the two concerns share a common feature: security systems built for today's environment may not remain adequate as computing capabilities change.
The payment industry therefore faces a layered security problem. Existing fraud must continue to be managed, artificial intelligence-enabled attacks must be anticipated and future encryption threats must be addressed before they become practical.
The emergence of autonomous artificial intelligence in cyber operations makes this more urgent because security failures can propagate quickly across interconnected systems. Visa's response reflects a broader recognition that cybersecurity is entering an era in which machines increasingly defend against machines.
The resulting challenge is not simply building better artificial intelligence. It is establishing sufficient controls around increasingly capable systems so that greater automation improves security without creating new vulnerabilities that attackers can exploit.
(Source:www.thedailystar.net)
