The emerging concerns over artificial intelligence's (AI) ability to ensure data confidentiality and privacy may prompt many businesses to reconsider their dependency on cloud-based AI systems, opting for on-premise servers instead. This is a significant u-turn from the initially popular trend towards cloud computing. It throws into question the often-held belief that the era of on-site servers is over.

"Most people were under the belief that the days of on-premise servers had passed, nonetheless, now we're witnessing a possible resurgence," commented Mark Presnell, the Managing Director of Convergence, an Auckland-based eCommerce integration company. He voiced concerns over the reliance of companies on the cloud to safeguard sensitive data. "The abilities of AI are impressive, but they also create unease about entrusting crucial information to the cloud. Companies using AI on-site may not desire to share the insights, ideas, and plans they generate on an open platform, which is the usual practice nowadays."

AI's main function, as Presnell highlights, is as a giant recall tool, contrary to regularly held misconceptions about its 'intelligence'. "Consider AI as an immense library. AI can locate specific information in seconds, a task that might take a human days. However, this is recall, not creation. Generative AI, which produces outputs such as text or images, exclusively relies on pre-existing data. It's not inventing something new; rather, it's rearranging known elements in innovative ways."

Although AI can create artworks or compose coherent text that appears unique, these are based on existing patterns and information. Presnell underscores that, "AI lacks original thought. It's a potent tool for processing and recalling information but not for originating it."

This understanding of AI as primarily a recall tool has implications for businesses and research institutions. Presnell identifies three critical areas of focus. First, companies, particularly those in customer service and retail, should concentrate on developing their internal knowledge bases, utilising AI's recall abilities to enhance customer interactions. Second, businesses should invest in comprehending AI applications specific to their industries, staying abreast of trends and considering a return to more robust on-site systems to handle the extensive AI data.

Finally, the issue of data ownership is crucial. "When you input information into cloud-based AI services, it usually becomes their property. Businesses must be deeply aware of this and contemplate the long-term consequences," warned Presnell.

In conclusion, Presnell believes the shift back to on-premise AI solutions reflects a deeper understanding of AI's capabilities and limitations, coupled with increased attention to data security and privacy in a progressively digital age."