COGNIZANT & IBM
“ The most important concern from my side is that organisations are rushing to use the LLM without having robust data security practices ”
MR . PAWAN GUPTA CYBER SECURITY GLOBAL OFFERING HEAD , CIS SECURITY AT COGNIZANT organisations and subject matter experts in this space allow both Mr . Gupta and Dr . Muppidi to visualise oncoming threats in the future , which – in some cases – may come from unknowing client situations themselves .
“ The most important concern from my side is that organisations are rushing to use large language models ( LLMs ) without having robust data security practices . So , data leakage – data security – is the primary concern across all industries as far as AI is concerned ,” says Mr . Gupta .
Dr . Muppidi says : “ There is no AI without IA ( Information Architecture ).
“ The secondary concern is the impact on supply chains . There is enormous usage of AI in code development and applications , but no one is in a position right now to say that the code has exposed them to back doors or maybe even malicious content – this could become part of an application and will be shipped across .”
He adds : “ While there will be further developments to secure Gen AI , we must recognise the basics as we adopt AI , i . e ., a programme for compliance and governance and securing the infrastructure and the applications using Gen AI .”
The two experts ’ comments make it clear that governments and industries have collaborative roles to play in making Gen AI suitable for everyday use and addressing the risks associated with doing so .
The pair put regulation high on the priority list as more consideration is to be given to the ethical use of AI and , in fact , the safe use of AI .
“ With the cloud journey , the industry first adopted the shared responsibility model ,” Mr . Gupta explains . “ With AI right now , we haven ’ t defined any shared responsibility model , and there are three elements that we currently speak about : the data , the algorithm , and then the application of it .”
With multiple cybersecurity elements and complexities in Gen AI , regulations are vital to ensuring the development of LLMs for high-profile functions . They will also standardise how such technology is used in variable environments . cybermagazine . com 93