Turkish Defense Firm Havelsan Develops AI Assistant

Last Updated on January 17, 2024 by SPN Editor

Havelsan, a prominent Turkish defense firm, has recently unveiled its cutting-edge AI assistant, MAIN (Multifunctional Artificial Intelligence Network). Scheduled for public release on February 1, MAIN specializes in natural language processing (NLP) and operates on the indigenous GPT model (Generative Pre-Trained Transformer).

This AI Assistant is based on a Generative Pre-Trained Transformer (GPT) model, which is adept at a variety of text-related tasks and is ready to delve into multimedia content. MAIN goes beyond the typical capabilities of a virtual assistant, with features to condense text and produce computer code. In developing MAIN, Havelsan has placed a strong emphasis on data privacy, using a unique hardware kit to avert data breaches. 

Developed by Havelsan’s Information and Communication Technologies team, led by Osman Kavaf, the AI Assistant sets itself apart with its emphasis on data security and versatility. Kavaf explained that the project began about a year ago in anticipation of OpenAI’s ChatGPT, to develop specific algorithms for text and multimedia.

MAIN, as the indigenous AI Assistant is named, boasts advanced features, including text summarization and the ability to generate computer code. This versatility allows users to input different parameters for iterative output, making it a multifunctional tool.

To ensure the security of data, Havelsan has implemented a unique approach, requiring a proprietary hardware kit for access. Kavaf described this kit as a specialized, closed-box device designed to prevent data leaks, emphasizing the company’s commitment to handling sensitive and confidential information responsibly.

In terms of accessibility, Havelsan is exploring the potential integration of cloud systems for remote access to the AI model in the future. However, Kavaf noted that, for now, this option is not prioritized due to the sensitive nature of the data involved.

The development team at Havelsan comprises 20 individuals, primarily academics, with plans to gradually expand the team. Additionally, the company aims to collaborate with engineering students from partner universities. Kavaf expressed the company’s dedication to data security and privacy and emphasized the ongoing efforts to establish mechanisms for public service in future versions.

A major AI event is in the pipeline for the launch day on February 1, where Havelsan plans to showcase the capabilities of the MAIN model. Kavaf envisions future stages of development to include enhancements in image and audio processing, positioning MAIN as a comprehensive solution that could revolutionize various aspects of daily life.

As the world becomes more aware of the potential of AI Assistant, MAIN’s initial release will be text-focused, with a strategic plan to incorporate image and audio processing in subsequent updates.

How does Havelsan plan to use AI Assistants in defense applications?

The Company is an expert in C4ISR (Command, Control, Communications, Computers, Intelligence, Surveillance, and Reconnaissance). The company places a special emphasis on various domains, including Command & Control Systems, Homeland Security, Simulation Technologies, maintenance, and logistic support for avionic systems and simulators, along with e-Government applications.

In the defense sector, Havelsan extends its offerings to encompass enterprise resource planning (ERP) solutions, operation centers, and integrated border security systems tailored for both military and civil applications. Additionally, the company is actively engaged in providing cybersecurity solutions, cloud computing services, and artificial intelligence solutions.

While the specific applications of the AI Assistant MAIN model in defense-related contexts are not explicitly outlined, it is reasonable to infer that MAIN could be seamlessly integrated into these domains, augmenting Havelsan’s existing capabilities with advanced AI functionalities. The potential applications may include improving data analysis within C4ISR systems, enhancing simulation technologies, or fortifying cybersecurity measures.