No data is "truly" secure if it can be accessed. A lot of AI tools that are available are AIaaS - AI as a Service. It's hosted somewhere else. There are security controls, but there are also tradeoffs. You need to understand them and make decisions accordingly.
You can get closer to "truly secure" with an on-premise implementation of AI - a controlled environment, no third party access or retention of data, physically isolated, complete control over network connectivity, logging, monitoring, encryption, access management...
However, if you're ignoring consistent patch management, identity controls, inadequate monitoring - physical and virtual - security expertise, outdated hardware or software, and internal threats, it's more of an illusion of security. I'd offer that, in this case, it may be less secure than AIaaS, but if the AIaaS provider has the same problems and you don't know about it... Saving Changes...
Program Manager| HARPER SRLSanto Domingo / Distrito Nacional, Dominican Republic
Data security starts with governance. Organizations need clear policies on what data can be shared with AI tools, use approved platforms, control access, and ensure sensitive information is protected or anonymized. Technology is important, but user awareness and responsible AI practices are just as critical. Saving Changes...
Community Champion
