According to IBM, there are three broad classifications for AI:<\/p>\n\n\n\n
- \n
- Narrow AI, <\/strong>also known as ANI, <\/strong>is the only type that currently exists; this type is trained to complete one task from a small spectrum of tasks. A good example is a spam filter, which is trained to catch unwanted emails.<\/li>\n\n\n\n
- General AI, <\/strong>also known as AGI<\/strong>, is a hypothetical form of AI that would use cognition to complete a wide range of tasks, akin to a human. A theoretical example would be an AI in the role of a scientist, where it generates hypotheses, executes experiments, draws correlations and adapts its research based on results.<\/li>\n\n\n\n
- Super AI,<\/strong> also known as ASI<\/strong>, is another hypothetical form of AI that transcends the capabilities of human intelligence. It\u2019s so intangible that an example cannot be ascertained.<\/li>\n<\/ul>\n\n\n\n
Given these definitions, the AI we know and apply in today\u2019s engineering field falls beneath the Narrow AI definition.<\/p>\n\n\n\n
Predictive AI with CAD software design<\/strong><\/h2>\n\n\n\n
Beneath the broad category of Narrow AI, AI within engineering design can take several sub-forms. For example, predictive AI is reshaping how engineers derive great value from their computer-aided design (CAD) software.<\/p>\n\n\n\n
Specifically, predictive AI analyzes historical and real-time CAD data to anticipate user needs and workflows before providing input. This predictive capability is powered by large language models (LLMs), natural language processing (NLP), and large datasets \u2013 meaning, the AI is trained on natural language text and source codes to learn context, patterns and relationships. Machine learning algorithms help the AI determine common key words within commands, which are paired with a user\u2019s real-time input and recently used commands to understand intent. This analysis then yields context-aware suggestions.<\/p>\n\n\n\n
With the right CAD solution<\/a> in place then, engineers can use predictive AI features, such as command prediction, to accelerate modeling processes and reduce the likelihood of errors; as command prediction observes user behavior before making suggestions that optimize industrial machinery and components designs.<\/p>\n\n\n\n
![\"\"](\"https:\/\/blogs.sw.siemens.com\/wp-content\/uploads\/sites\/2\/2025\/09\/AI-UC-Gen-AI-Hydrogen-Visual_original-1024x576.png\")
<\/figure>\n\n\n\nGenerative AI with design<\/strong><\/h2>\n\n\n\n
In a similar vein, engineers can gain design value from the subfield of generative AI. Like predictive AI, generative AI analyzes large datasets for patterns — but instead of predicting user intent and making suggestions, it uses these patterned connections to create content, such as text, code, images, or even music.<\/p>\n\n\n\n
Specifically, as generative AI analyzes massive data sets for structures and patterns, it uses a complex algorithm known as a neural network to detect and interpret the patterns. Neural network architectures can vary; the Transformer model tends to be one of the most popular models for text-based generative AI, as it processes sequential data before generating new sequences that take the form of content (text). This content reflects the patterns the Transformer model ascertained from its dataset analysis. After generation, content is at the discretion of the user and can be implemented as desired.<\/p>\n\n\n\n
While we\u2019ve all heard stories of the popular generative AI platform, ChatGPT, in engineering design, generative AI is even more powerful. The right CAD software<\/a> provides generative design features that suggest lines of code; this code can optimize industrial machinery structures, weight, performance, appearance and several other design aspects. By designing and modeling the best possible product, engineers can shrink their design cycle times, expediting product development so that more products are introduced to the market sooner.<\/p>\n\n\n
- General AI, <\/strong>also known as AGI<\/strong>, is a hypothetical form of AI that would use cognition to complete a wide range of tasks, akin to a human. A theoretical example would be an AI in the role of a scientist, where it generates hypotheses, executes experiments, draws correlations and adapts its research based on results.<\/li>\n\n\n\n