Generative AI offers unprecedented capabilities to sift through massive amounts of data stored in a variety of formats, such as documents, tables, and charts, but also audio and video, leading to the development of so-called multi-modal AI platforms.
One of the generative AI applications developed by FedData Technology Solutions (FDTS), involves a chatbot-like system that answers questions on US Army regulations, we call it ArmyGPT. Trained on a large corpus of official US Army documents, ArmyGPT can answer any question on US Army regulations using natural language. A conservative estimate indicates that a 200 pages PDF can be embedded in the AI model in about 30 seconds.
FDTS has capabilities to deliver fully air-gapped and on-prem generative AI applications that do not require any interaction with external providers, i.e., OpenAI. This approach ensures two aspects of the cybersecurity triad: data confidentiality and integrity. In addition, FDTS’s extensive network of state-of-the-art hardware vendors such as Dell, NVIDIA, and Supermicro, ensures also data availability, since our solutions can be deployed on-premises using dedicated hardware. The latter aspect satisfies also data availability, therefore completing the cybersecurity triad.
Another key aspect of our AI solutioning approach is testing and validation of generative AI applications. It is paramount to ensure that the generative AI explores the appropriate dataset pertinent to the application at hand, and to avoid what in AI-jargon is referred to as “hallucinations”. FDTS provides this capability by limiting the range of answers that the AI model can provide.
Recent trends in generative AI are showing an unprecedented adoption of this technology on a broad spectrum of human activity; while applications in science and technology are leading the way, generative AI is making a significant impact in the medical field with “intelligent virtual doctors” applications capable of providing a “human-like interactions” via fine-tuned natural language processing (NLP) models. Other interesting areas are also adopting this technology, including finance, marketing, and even arts and graphics design.