The CIO's guide to generative artificial intelligence

Rennie Naidoo, professor at the Wits School of Business Sciences, Department of Information Systems, University of the Witwatersrand.

---

In the rapidly-evolving landscape of artificial intelligence, GenAI stands as a transformative force for creating new sources of business and social value.

According to a recent PwC report, 84% of chief information officers (CIOs) plan to adopt GenAI. For CIOs, navigating this new and complex terrain is both a strategic opportunity and challenge.

Drawing from a recent environmental scanning exercise, this quick guide aims to provide CIOs with strategic insights into harnessing the potential of GenAI, balancing innovation with ethical considerations, and aligning technology with value creation.

In their roles as futurists, CIOs and their strategy teams can also employ scanning techniques to continuously survey and analyse their environment to identify opportunities and threats tailored to match the needs of their narrower organisational context to navigate the GenAI landscape effectively.

What follows is my interpretation of key external data and recent trends in GenAI, and general implications for IT leaders.

Predictive AI is about anticipation and forecasting, whereas GenAI is more about creation and innovation, generating new outputs based on learned patterns and inputs.

In other words, predictive AI focuses on analysing historical data to identify patterns and make forecasts about the future, while GenAI is about creating new content, be it text, images, audio, or video.

GenAI presents a unique opportunity for CIOs to drive innovation and deliver significant business value aligned to broader strategic objectives.

In the foreseeable future, predictive AI will probably remain the primary go-to technology for many AI-based applications. CIOs should be aware that while GenAI is potentially transformative, it is currently only appropriate for some use cases and models. Moreover, there are use cases where predictive and GenAI can work together in various applications.

Given the rapid change and growing unpredictability in today's business landscape, adopting GenAI in enterprises demands a cautious approach.

A critical challenge in embracing GenAI is navigating through the marketing hype, often referred to as 'AI-washing', to achieve the evolving aspirations of the organisation.

CIOs must rigorously vet technology vendors to ensure their AI capabilities are genuine and substantiated. This involves understanding the specifics of the technology, conducting pilot tests, consulting with AI experts, and setting clear performance metrics and benchmarks.

A thorough vetting process is crucial in selecting AI solutions that genuinely offer value and align with the organisation's strategic objectives.

The rapid adoption of GenAI must be balanced with stringent ethical considerations. This includes developing ethical frameworks for AI, focusing on data privacy, minimising biases and ensuring transparency in AI processes.

Acceptable use policies should clearly define the purposes for which AI is used within the organisation, accompanied by regular employee training and awareness programmes. Continuous monitoring and evaluation of AI systems are essential to ensure they adhere to ethical standards and regulations.

The potential of these technologies to revolutionise business processes is immense, but so are the challenges they pose, particularly in terms of ethics, privacy and security.

For CIOs, the focus should be on implementing GenAI ethically, securely and in line with the organisation's core values. This means closely monitoring the development of AI systems to avoid inherent biases, ensuring transparency in AI decision-making, and establishing clear accountability for AI-driven outcomes.

The CIO's strategy should also reflect an awareness of the ethical, security and societal challenges posed by rapid advancements in AI.

Identifying strategically aligned and technically feasible use cases for GenAI across business functions is critical. For IT service desks, GenAI can automate ticket resolution and aid in knowledge base development. In software development, it can assist in code generation and bug identification.

Beyond IT, GenAI can be applied in marketing, human resources, customer support and finance. For example, in the South African finance sector, it can assist in fraud detection and risk assessment.

Use cases with a quicker time to value should be prioritised, especially for initial ventures into AI. Scaling these use cases requires pilot testing, cross-functional collaboration and continuous learning and adaptation.

At the same time, CIOs and IT leaders must understand and address misuse cases. For example, while AI can boost productivity and innovation, it can also lead to job displacement, particularly in automation-prone sectors.

Moreover, if trained on biased data sets, AI systems can perpetuate and amplify biases. This can lead to unfair or discriminatory outcomes. AI can also be used maliciously to enhance the capabilities of cyber attackers, making threats like phishing, identity theft and network attacks more sophisticated and harder to detect.

Recent reports praise the remarkable efficiency of tools like GitHub Copilot and ChatGPT, which have been trained on billions of lines of code. These AI platforms function like advanced autocomplete systems, accurately predicting code lines and significantly reducing coding time.

These tools not only save time but also enhance job satisfaction, with many software developers reporting more fulfilling work experiences. However, their reported limitations − notably the occasional production of irrelevant code or 'hallucination' − underscore the continued necessity of skilled developers for oversight.

Looking ahead, AI's role is set to expand beyond just writing code to include architectural decisions, code reviews and project management. This evolution promises to flatten the learning curve for novices, while empowering experienced developers with the freedom to innovate and focus on strategic development areas, thereby driving overall productivity and creativity in IT departments.

Upskilling and training employees to leverage GenAI tools effectively are vital. This involves investing in continuous education, fostering collaborative innovation and implementing agile methodologies.

Ensuring AI initiatives are aligned with the organisation's strategic goals and fostering a culture that values learning and innovation is crucial to maximising the technology's benefits across the workforce.

Regulatory scrutiny and compliance become increasingly important as AI technologies become more integrated into business operations.

CIOs must navigate this landscape by developing transparent AI operations, addressing unsubstantiated claims, ensuring privacy and data protection, tackling biases, and engaging in policy discussions and advocacy.

Ethical AI frameworks should go beyond legal compliance, focusing on broader societal impacts. There is no need for a complete overhaul of governance and testing for GenAI, as the rigorous testing, validation and constant monitoring processes for predictive AI can often suffice.

Setting clear goals and selecting the right AI solutions to ensure tangible business outcomes will be a major paradoxical challenge facing many CIOs.

Measuring the return on investment and productivity gains from GenAI while concurrently assessing and managing potential risks is essential but complex in practice. This dual focus involves not only assessing the impact on customer satisfaction, service availability and operational costs, but also identifying and mitigating risks, such as data breaches, ethical missteps, or unintended biases in AI applications.

This approach helps in balancing innovation with accountability, ensuring productivity gains from AI do not come at the cost of increased vulnerability or ethical compromises. Again, experts suggest the processes for managing GenAI risk will be similar to those of predictive AI.

CIOs must keep up with rapid advancements in AI technology and be prepared to adapt their strategies as new solutions emerge. This involves staying informed, assessing the relevance and readiness of new technologies, managing risks and ensuring ethical AI practices.

Building a scalable AI infrastructure and maintaining ongoing vendor relationships are also crucial for future-proofing AI strategies. GenAI presents a unique opportunity for CIOs to drive innovation and deliver significant business value aligned to broader strategic objectives.

However, this requires a strategic approach that prioritises ethical considerations, aligns with business objectives and adapts to the rapidly-changing AI landscape.

While focusing on fit-for-purpose solutions and fostering a culture of continuous learning and innovation can unlock the enormous potential of GenAI, CIOs must also be vigilant about the challenges it poses, including managing data privacy and cyber security concerns, mitigating bias in AI algorithms and addressing the potential for massive job displacement.

These challenges necessitate a balanced approach to ensure the integration of GenAI into business processes enhances rather than compromises organisational integrity and public trust to keep organisations competitive in the dynamic and rapidly-evolving AI-driven world.

Despite the inherent subjectivity, employing environmental scanning techniques through a collective intelligence system to identify opportunities and challenges can offer CIOs valuable insights for thoughtfully shaping and integrating GenAI into their business strategies.

* This article draws on ongoing research at Wits University, addressing critical areas for successful GenAI integration in enterprise environments and IT units.