AI for businesses

The Differences Between Open and Closed AI

Earlier this year, Matt Riley, Director of Transformation and Security at Sharp UK, wrote a blog which explored the two types of AI – generative and semantic search. In this blog, Matt expands on another concept within AI – Open AI and Closed AI and the impact of that on businesses, revealing its potential to both hinder and help them.

Artificial Intelligence (AI) represents a transformative technological leap, empowering machines to simulate human intelligence. AI enables systems to learn, reason, and make decisions. Its applications span industries, revolutionising processes, enhancing efficiency, and shaping the future of technology and business.

Open AI

Most people will have heard of ChatGPT – this is an example of an OpenAI model. An OpenAI model, often known as Public AI, is a type of AI model that is publicly accessible and can be modified by anyone.  

OpenAI models are designed to be more flexible and adaptable, capable of learning and evolving over time. They are trained on publicly available data from across the internet, such as text articles, images, and videos. OpenAI models have the source code openly shared so that people are encouraged to voluntarily improve its design and function.

However, OpenAI models can pose a severe data security risk. Data input into an OpenAI model can be accessible by anyone seeking that data (such as trade secrets and sensitive personal information). Confidential, personal, and/or personally sensitive data should never be input into an OpenAI engine. As such, careful consideration and responsible usage are necessary to mitigate potential risks associated with the accessibility of OpenAI models.

Closed AI  

Closed AI is a type of artificial intelligence model that is not publicly accessible although they may be trained on publicly available data. An example of a closed AI model is Bing Chat Enterprise. Unlike OpenAI, Closed AI models prioritise data security and confidentiality, maintaining strict control over internal data access and usage to safeguard sensitive information. This distinction ensures a higher level of data privacy, particularly concerning personally identifiable or confidential data.

It is recommended that businesses only use Closed AI systems to strengthen their defences against cyber threats and phishing attacks, ensuring robust protection of sensitive data and bolstering overall cybersecurity measures.

Why this is important now

Microsoft introduced their AI tool, Microsoft CoPilot, accessible within Microsoft Edge, offering two variants: Bing Chat (Open AI) and Bing Chat Enterprise (Closed AI).  

Using Bing Chat Enterprise within Microsoft Edge is crucial, identified by a green "protected" icon at the top of the page. Opting for the Closed AI version is essential, mitigating vulnerabilities that could expose business-related information to potential hackers. Leveraging the Closed AI model in Bing Chat Enterprise enhances data security, curtailing the risk of cyber threats and safeguarding sensitive business data.

Protecting your business from cyber criminals is crucial in today's interconnected digital landscape. With increasing reliance on technology and data, the stakes are higher than ever. Cyber attacks not only jeopardise sensitive information but can also pose severe financial and reputational risks. 

The evolving sophistication of cyber threats demands robust security measures including building a security aware culture within your organisation. Using only closed AI is a small but key step to protecting your business.  

Find out more about the responsible use of AI in the workplace