Debunking Generative AI Myths for CIOs

  • juillet 24, 2024
young boy smiling at robot toy

Topping the priority list for many IT leaders, Generative AI (GenAI) is poised to revolutionize the technical landscape.

According to Gartner®, “by 2026, more than 80% of enterprises will have used generative artificial intelligence application programming interfaces (APIs) or models and/or deployed GenAI-enabled applications in production environments, up from less than 5% in 20231.”.

Despite its potential, misconceptions persist, creating hesitation among decision-makers. Let's navigate through these myths with an analytical lens, debunking falsehoods and illuminating the true potential of GenAI.

Myth 1: GenAI is a plug-and-play solution

The reality: Implementing GenAI can often be complex, requiring substantial planning, customization and ongoing management. Successful deployment involves understanding specific business needs, integrating with existing IT services and continuously refining AI models. For example, virtual assistants deployed in customer service must be tailored to unique business protocols. In a future scenario, a retail company could implement GenAI for customer support, requiring ongoing adjustments to align the AI with evolving customer service standards.

Myth 2: GenAI is limited to generating text, such as articles or chat responses

The reality: GenAI’s capabilities extend far beyond this. Recent advancements, like OpenAI's GPT-4O, have introduced multimodal functionalities, enabling real-time speech recognition, visual understanding, and emotional expression. This evolution can transform applications across sectors — from developing advanced diagnostic tools in healthcare to creating immersive and empathetic customer service experiences in retail. Imagine how a manufacturing company could use GenAI for visual quality inspections, reducing defects and improving production efficiency through real-time image analysis and anomaly detection.

Myth 3: Generative AI eliminates the need for human oversight

The reality: Human oversight is critical to monitor performance, ensure ethical use and make necessary adjustments. AI models can generate outputs that may not always align with business goals or ethical standards. Organizations are establishing AI ethics boards to oversee AI projects, ensuring alignment with ethical guidelines and societal values. Continuous monitoring helps in catching biases and inaccuracies that could lead to unintended consequences.

Myth 4: Generative AI guarantees immediate ROI

The reality: The benefits of GenAI often unfold over time. Initial stages may involve significant investment in terms of time, resources, and training. The true value becomes clear as the AI system matures and integrates deeply into business processes. We have emphatically observed that companies seeing the most significant returns from AI are those that have patiently iterated on their AI models and strategically aligned AI initiatives with long-term business objectives.

Myth 5: Bigger AI models always yield better results

The reality: Bigger isn't always better. Clients often ask us to train large language models (LLMs) for their use case. However, many instances don’t require new LLMs. Smaller, fine-tuned models can outperform larger ones when tailored to specific industry jargon and user interactions. The effectiveness of an AI model depends on the quality and relevance of the training data, not just the model's size. Tailoring smaller, domain-specific models often results in better performance for applications.

Myth 6: GenAI cannot be integrated with legacy systems

The reality: Modern GenAI services are designed to be interoperable across various platforms. For example, NTT DATA’s mainframe modernization services and UniKix platform can enable seamless integration of GenAI with existing IT infrastructures, ensuring businesses can use AI capabilities without overhauling their entire system. Carrying through with an example, it is therefore possible for a healthcare provider to successfully integrated GenAI into a decades-old patient management system, thus enhancing patient care.

Myth 7: GenAI systems are inherently secure by design

The reality: AI systems are vulnerable to various security threats, including data breaches and adversarial attacks. And given the newness of the technology, we’ll continue seeing Implementing robust security measures at every stage of AI development and deployment is essential.

Myth 8: Generative AI can fully replace traditional applications

The reality: GenAI is most effective when it complements and enhances traditional applications, not replaces them. It can automate repetitive tasks, provide insights, and enhance user experiences, but core functionalities of traditional applications often remain vital. In future industry scenarios, GenAI can enhance capabilities by providing advanced data analytics and predictive insights, without entirely replacing the underlying systems.

Myth 9: Generative AI deployment is a one-time project

The reality: AI deployment is an ongoing process involving continuous learning, updating, and refining. Regular updates and retraining are necessary to keep the AI models relevant and effective as new data and technologies emerge. Consider the case of retail companies continuously updating their recommendation engines based on new consumer behavior data, ensuring their AI systems stay effective and competitive.

Myth 10: GenAI technology is still in its infancy and not ready for enterprise-level applications

The reality: Rapid advancements and successful implementations across various industries prove otherwise. Recent updates by Google and OpenAI prove the maturity and robustness of GenAI models. As an example, pharmaceutical companies can utilize GenAI for drug discovery, drastically reducing the time required for research and development phases.

Ethical Considerations

As GenAI and AI more broadly becomes more prevalent, organizations must prioritize ethical considerations. This includes setting up clear guidelines for responsible AI development, addressing potential biases and ensuring transparency and accountability in AI decision-making processes.

Implementing GenAI: a roadmap

  1. Define business aims and use cases.
  2. Assess existing IT infrastructure and data readiness.
  3. Select the appropriate AI model and platform.
  4. Develop a plan for integration and deployment.
  5. Implement robust security and ethical measures.
  6. Continuously monitor, update, and refine the AI system.

Conclusion

Generative AI holds transformative potential for application services, but it is crucial to approach its implementation with a clear understanding of its capabilities, limitations and ethical considerations. By debunking these myths and following a strategic roadmap, CIOs and CTOs can leverage GenAI to drive innovation, efficiency and competitive advantage while prioritizing responsible and secure AI development.

Explore GenAI opportunities tailored to your business needs. Visit our Application Services page for more insights.

Sources

1Gartner Press Release, Gartner Says More Than 80% of Enterprises Will Have Used Generative AI APIs or Deployed Generative AI-Enabled Applications by 2026, October 11, 2023. https://www.gartner.com/en/newsroom/press-releases/2023-10-11-gartner-says-more-than-80-percent-of-enterprises-will-have-used-generative-ai-apis-or-deployed-generative-ai-enabled-applications-by-2026 GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.”

Subscribe to our blog

ribbon-logo-dark
Shakti
Shakti Pradhan
Shakti has broad range of expertise in IT service delivery, service excellence, digital transformation, infrastructure transformation and automation programs for large clients. He is passionate about future technologies, productive approach and result-oriented strategies. In the client growth office, he is leading the initiatives in NTT DATA’s Digital Application Services (DAS) offering including application management, application modernization, quality engineering and assurance, performance monitoring and observability, security and portfolio management.
1437529-Drew-Gregory-headshot.jpg
Drew Gregory

Drew’s previous experiences across a spectrum of IT services including multiple levels in leadership have enabled him to holistically drive connectivity between business strategies and IT solutions. Today, he leads NTT DATA’s Digital Application Services (DAS) offering including application management, application modernization, quality engineering and assurance, performance monitoring and observability, security, and portfolio management.

Related Blog Posts