In Australia, OpenAI’s artificial intelligence agent gained unauthorised access to the state medical statistics portal “Medicare” while conducting research. After the incident, the government decided to introduce mandatory reporting of security incidents for artificial intelligence companies and to create a separate oversight mechanism. According to the company’s investigation, there is no evidence that patient data was accessed.
What dangers does granting artificial intelligence free access to certain systems create? Can citizens’ medical, banking and personal data be leaked through artificial intelligence in Azerbaijan?
Eshgin Nurushov, technology director of the company “ITStation”, told Globalinfo.az that when artificial intelligence is given access to any system, the security issue must be considered first of all:
“Because artificial intelligence can be given the ability not only to read data, but also to change, transfer and delete it. In this case, the exploitation of gaps in the system leads to serious problems. For example, medical data, bank accounts and personal information may be put at risk. Artificial intelligence is not a hacker in itself, but those who use it can turn this technology into a tool for cyberattacks.”
The expert said that the existence of such risks in Azerbaijan cannot be ruled out either:
“Because state institutions, banks and medical facilities store large volumes of citizen data. If uncontrolled access to these systems is obtained, data leaks are possible. In particular, the spread of medical data can affect people’s private lives, while the seizure of bank data can cause direct financial damage.
Eshgin Nurushov
The main issue here is what powers artificial intelligence has. Giving it free access to all data is not the right approach. In every system, access rights must be limited, user activity must be monitored and security checks must be carried out. Sensitive data must also be encrypted, and its transfer and sharing must be kept under control. It must be determined in advance which operations artificial intelligence can carry out independently. In particular, operations such as deleting, modifying and transmitting data must not be carried out without human approval.”
Nurushov noted that one of the main dangers in the future will be people’s excessive reliance on technology:
“Artificial intelligence can work fast, but that does not mean it always makes correct and safe decisions. Therefore, alongside the development of technology, cybersecurity measures must also be strengthened. The main principle is that artificial intelligence should be given only as much data and authority as needed, and all processes should be kept under control.”
Laman Imran
Globalinfo.az
Read also
