What is generative AI?

Generative artificial intelligence (AI) has already significantly impacted the ways we interact with technology, solve problems and perform day-to-day tasks.

Awareness of the risks, challenges and best ways to integrate the technology into your organization is key to success.

What is generative AI?

Generative AI refers to a class of AI models that can generate new and original content, rather than just analyzing existing data. Unlike discriminative models (which distinguish between different categories), generative AI can identify underlying patterns and structure data to produce novel outputs that mimic that data.

For example, discriminative AI can look at a thousand pictures of cars and motorbikes and learn to tell the difference. Generative AI, however, can look at the same images and then create a brand new image of a car or a motorbike that never existed before.

Generative AI also has key differences and limitations in comparison to agentic AI. For example, unlike agentic AI, generative AI responds to human prompts to generate its content. Agentic AI, on the other hand, can autonomously produce content without the need for prompts.

The benefits of using generative AI

Generative AI has a wide range of applications:

• Synthetic data-generation: When data is scarce, sensitive or hard to obtain, generative AI models can be used to create realistic, privacy-preserving datasets for training other models. This is particularly helpful in industries with strict data privacy regulations, such as healthcare and finance.
• Creative and predictive analytics: Generative AI models are being used to create new product designs and even to forecast future trends by simulating various scenarios.
• Data augmentation: Generative AI can significantly expand the size and diversity of a dataset by generating new variations of the existing data.
• Text analytics: By far the most common use of generative AI is its powerful ability to summarize large text documents, extract key insights and generate reports. For example, an individual could use an LLM to automatically generate a summary of customer feedback.

Challenges and ethical considerations

Data professionals must address the significant challenges and ethical considerations that come with the use of generative AI. For example:

• Data privacy and security: The vast amounts of data used to train the models raise concerns about potential memorization and leakage of private information.
• Bias: Building and using a responsible AI should be a key consideration. If the data contains any biases (e.g., gender, racial, or cultural), the model will learn and amplify these. This can lead to unfair and harmful outputs, meaning their use requires careful data curation and ongoing monitoring.
• Intellectual property: The use of copyrighted material in training data and the attribution of ownership for AI-generated content is a complex legal and ethical issue that continues to be debated.

Applications of generative AI

How Bayer is streamlining agronomic data analysis

Bayer is tackling the immense challenge of helping farmers optimize crop yields. The company, a leader in the life sciences and agricultural sectors, is tasked with analyzing a vast amount of complex and unstructured data. This includes agronomy reports, weather patterns, soil data and scientific research. Traditionally, a team of agronomists and data scientists would manually go through these data sets to provide recommendations.

Now, Bayer is developing a generative AI-powered assistant to streamline this process. This AI model is designed to act as a conversational interface for a massive internal data knowledge base. Agronomists will be able to ask questions such as: “What are the key agronomic challenges for corn in a drought-prone region of the US Midwest this season?”

"With its ability to improve the lives of millions, AI is a game-changer. It is a collaborative effort – made possible by the talent and dedication of teams across Bayer’s divisions and IT, as well as partnerships with leading technology companies and institutions. We have only begun to tap the transformative power of AI, and the best is yet to come," says Bijoy Sagar, Bayer's chief information and digital officer.

How Morgan Stanley integrated generative AI for knowledge management

Morgan Stanley is integrating generative AI internally to create a more efficient knowledge management system. The financial services giant partnered with OpenAI to build a custom solution that can analyze the company’s vast amount of market data and analyst reports.

“We’re sitting on a massive amount of intellectual property,” says Jeff McMillan, head of analytics and data at Morgan Stanley. “The challenge for our advisors was always finding the right piece of information at the right time. They’re constantly dealing with client questions that require them to pull data from multiple sources.”

Now, the custom-built AI assistant acts as a powerful search engine and helpful summarization tool. Rather than sifting through thousands of documents manually, advisors and employees can ask natural language questions and receive a concise answer with the relevant documents cited.

How Toyota is improving business intelligence (BI) with generative AI

Toyota North America (TMNA) faced a common challenge for large enterprises: data silos. Within the company, different departments, including manufacturing, sales and supply chain had their own fragmented data systems. This made it difficult to get a unified view of the business.

To tackle this, TMNA initiated a project to create a unified data lakehouse to implement “generate BI” tools. Their initiative includes:

• Natural language queries: Without needing a background in SQL or data visualization, individuals can simply ask questions in plain language, such as “Show me sales trends for the Corolla model in California over the last two years compared to the national average."
• Automated report generation: Their model can take a set of KIPs and generate a full business report, including graphs and a narrative summary.
• Predictive forecasting: It can also analyze historical sales to generate multiple plausible future scenarios to help decision-makers test different strategies.

The future of generative AI

The continued development of generative AI promises an intuitive future for data analytics. Generative AI will most likely move away from becoming a siloed tool, toward a foundational layer that permeates most aspects of the data ecosystem. 

It is also entirely plausible that the next generation of generative AI will be multimodal, able to seamlessly understand and generate content. 

Beyond multimodality, agentic AI is a major emerging trend. According to a 2024 report from Deloitte, generative AI will also move from being a simple “co-pilot” to an independent, task-oriented data agent. These autonomous systems can break down a complex problem into a series of steps, and access and utilize different tools such as databases, APIs and data visualization software.

Lastly, the tension between data utilization and privacy will be a major focus in the years ahead. Generative AI will play a critical role in this by becoming the primary engine for creating privacy-preserving synthetic data. As regulations like GDPR become more stringent, organizations will intranasally rely on generative models to create statistically representative, non-identifiable datasets for training and testing.

Pioneering the Next Era of Intelligence

Join All Access: AI, Data & Analytics 2025. A free webinar series streaming live November 4, 2025, designed to guide you in integrating AI effectively.

Learn from industry experts on identifying opportunities, managing risks, and making strategic AI and data investments. Key themes include Decision Intelligence, AI and Data Governance, and Scaling GenAI and ML Operations.

Learn More