Nvidia, the giant computing firm that offers AI solutions, has recently developed ‘NeMo Guardrails’. This open-source software offers solutions to some issues which AI technology currently faces in the field. Primarily, NeMe offers to limit the amount of inaccurate and random information generated by AI Chatbots like OpenAI’s ChatGPT and Microsoft’s Bing Chat.
NeMo Guardrails helps developers instate guardrails, aimed to keep conversations between the user and the chatbot on topic and to ensure that the AI tools provide accurate information. Moreover, this software provides a safeguard that prevents chatbot tools from connecting to any third-party applications that have not been authorized which could be a potential security risk.
The way NeMe works is by deploying a second logic learning model that cross-references the information provided by the first one. In this way, the accuracy of the responses is increased and the AI system is not easily provoked by stimuli.
However, the software does not take into account instances of bias and deception, as research done by the Center of AI and Digital Policy (at the Federal Trade Commission) has indicated.
Nvidia is a technology company based in the US, while they began as a GPU maker and focused primarily on graphical solutions, they made an effort to pivot their attention to AI solutions. Since this pivot, Nvidia is taking advantages of AI through the development of software and other AI solutions, which has made them financially profitable.
In addition, this development also comes as a response to the calling of tech firms such as Steve Wozniak and Elon Musk for a pause on AI development. Even though this won’t satisfy their calling, it does try to aid in the improvement of the technology.
