On September 12, 2024, Oprah Winfrey hosted a prime-time television special on AI, featuring Sam Altman, Bill Gates, and other prominent figures. Nothing captures the zeitgeist quite like a television special with Oprah, which tends to draw an audience of 10-15 million viewers.

The launch of ChatGPT 3.5 in November 2022 marked the dawn of the Age of AI. Since then, the industry has dominated headlines. Tech giants have poured billions into securing their place at the forefront, while companies across every sector are exploring AI to unlock breakthrough productivity gains. Meanwhile, policymakers scramble to provide the necessary guardrails, and tens of thousands of startups are flooding the market with AI-driven solutions.

In this context, Samudra members have come together to explore the potential impact of generative AI on organizations and communities. Samudra is an “ecosystem of communities of purpose” where leaders from diverse backgrounds collaborate to address complex, intersectional challenges. Participants include board members, CEOs, business and technology executives, as well as thought leaders from various disciplines.

As generative AI becomes more accessible to consumers and integrates into devices like smartphones, we can expect an explosion of new applications and use cases.

Below is a summary of the key insights from recent discussions. Samudra operates under “Chatham House Rules,” which does not allow perspectives and statements to be attributed to individual members. However, we are happy to share some high-level insights based on conversations and qualitative research:

1. AI investments will continue despite the “trough of disillusionment.”
Even as we go through market disillusionment and pullback, investments in AI will persist for years due to several driving factors: (1) the unwillingness of tech giants to relinquish market dominance to competitors, (2) the continued interest from companies who see benefits by deploying AI, and (3) the potential for a global AI arms race especially between China and the US compelling governments to aggressively pursue this technology.

The nature of AI investments will evolve, with rapid and often unpredictable shifts. For example, venture capital has already moved quickly from investing in AI infrastructure to AI applications, and now to AI “agents.” As generative AI becomes more accessible to consumers and integrates into devices like smartphones, we can expect an explosion of new applications and use cases.

The bigger, unaddressed question is whether current governance structures are sufficient. We cannot rely on public policy to keep pace, even if our governments are willing to regulate AI.

Fundamentally, generative AI is seen as a foundational platform shift similar to the transformative impact of mobile phones and the Internet with the potential to create deep, widespread changes across all industries and sectors of society. As such, significant investments in this space will continue.

2. Proper AI governance is a major concern at the top
In a recent poll among Samudra board members, AI emerged as one of three top priorities, with special emphasis on associated risks and governance. For example, there is concern about the potential liabilities companies could face if AI delivers a faulty response to a patient or causes financial harm to a customer.

The bigger, unaddressed question is whether current governance structures are sufficient. We cannot rely on public policy to keep pace, even if our governments are willing to regulate AI. Moreover, most corporate governance frameworks are not designed for such rapid change. There is a clear need to tightly integrate informed governance with product and service development and delivery.

This presents an opportunity for companies to join ecosystems where they can pool data for mutual benefit or to form partnerships where they exchange their data for AI expertise and development.

3. Many organizations are assembling internal platforms to develop AI-powered applications.
Companies are leveraging AI for business value in three main ways. The first is partnering directly with AI providers (e.g., OpenAI, Nvidia) to develop proprietary large language models (LLMs) for competitive advantage. This high-risk, high-reward strategy is feasible only for companies with deep financial resources and access to rare technical talent.

The second approach is to leverage the AI capabilities that come bundled with various cloud software providers. This is relatively easy to implement and can quickly offer limited productivity gains (e.g., instant document summaries, meeting transcripts, etc.), but experience shows that the costs are still too high compared to the value. Worse still, these gains fail to differentiate the business, as any company can subscribe to the same services.

The third option is assembling an internal platform that applies external LLMs to proprietary corporate data to create business-specific applications that provide a competitive edge. We see more organizations adopting this approach, with some vendors offering ready-made platforms to accelerate this process rather than building from scratch.

4. Data management with unstructured data will be essential for deriving value.
The technology leaders within Samudra such as Chief Information, Data, and Digital Officers consistently emphasize that they must manage and govern their unstructured data correctly in order to capitalize on the promise of AI. They refer to the entirety of documents and content assets in an organization spreadsheets, images, videos, contracts, architecture diagrams, formulas, project plans, resumes, customer feedback, and chats, just to name a few.

However, not all unstructured data hold value for the business. The challenge is identifying which data are unique and valuable to the company and then using AI to capitalize on that.

A corresponding insight is that not every organization has access to the volume and diversity of proprietary data necessary to gain a competitive edge. This presents an opportunity for companies to join ecosystems where they can pool data for mutual benefit or to form partnerships where they exchange their data for AI expertise and development.

5. ROI will hinge on enabling multiple, differentiating use cases over time.
Just as the Internet and mobile technology have transformed how we work and live our lives, we believe that the return on investment from AI will stem from enabling multiple use cases over time. This requires building a flexible AI-enabled technology platform to support multiple flavors of LLMs, AI applications, emerging devices, and more.

The best place to start is to identify what business challenges, customer needs, or new capabilities will benefit the organization most. In other words, don’t start with the technology; start with the business case. With this in mind, many organizations have established cross-functional “AI Governance Councils” with representatives from various departments (including legal, HR, finance, IT, and the business) to collectively assess opportunities, manage risks, and guide AI initiatives.

Leaders must uphold strong ethical standards in the deployment and use of AI.

Meanwhile, various vendors and startups are working hard to develop AI-based applications tailored to specific industries and use cases. Some examples include contract review for the legal industry, fraud detection in financial services, and creating voice-over videos in marketing.

6. Consider that Generative AI has emerging “superpowers” that you can leverage at scale.
While we consider business-differentiating use cases, it’s helpful to think of generative AI as having distinct “superpowers” we’re discovering over time. For example:

• Finding Information: AI can revolutionize search functions by quickly retrieving and synthesizing relevant data, potentially disrupting traditional search engines.

• Analyzing Information: AI can summarize content, answer questions about documents, and perform other analytical tasks, enhancing our ability to process and understand information.

• Creating Information: AI can generate a wide range of content, including images, videos, text, and code, enabling new forms of creative and technical expression.

• Performing Tasks: AI-driven automation and intelligent agents can handle repetitive tasks, streamline workflows, and support various business processes.

• Helping People: AI can assist with training, learning, and even coaching, providing personalized support and enhancing human development.

These superpowers become even more impressive when applied at scale. Analyzing a single document is relatively simple, but analyzing 1,000 legal documents simultaneously is a significant challenge. Similarly, generating a few images is one thing, but producing hundreds of “good enough” images quickly ones that your creative team can refine into final products is another level of capability and value.

7. Change management and adoption will be essential for realizing value.
As with any technology, adoption and effective use are crucial for success. AI presents unique challenges in driving change and increasing utilization within organizations.

AI clearly has the potential to displace workers, possibly eliminating entire job families. We’re already seeing early signals where businesses let go of lower-performing employees to offset the high costs associated with implementing AI, or reduce hiring for open positions, expecting AI to boost productivity.

This puts significant structural and cultural change pressure on companies. Depending on how this change is managed, employees may perceive AI as a threat or an ally, which can accelerate or inhibit adoption and value. At any rate, the impact of AI on the workplace environment will be massive, politically charged, and will challenge leaders in new and complex ways.

Studies show that employees often use AI in unsanctioned ways, frequently utilizing publicly available LLMs and applications. Given the rapid pace of change, many organizations are still in the process of developing internal AI policies, security protocols, and usability cases. Leaders need to address how to manage the transition from personal to professional AI use and drive the necessary behavioral changes.

8. We need more human(e) leadership to navigate us through this time.
The role of leadership in effectively navigating through this period cannot be overemphasized. While awareness of AI as a technology is essential, the required qualities to leverage the new opportunities are deeply human and even humane.

Critical leadership traits include:

• Trust: Build and maintain trust within the organization. Build AI systems and applications that increase rather than decrease trust.

• Empathy: Understand and address the concerns and emotions of employees as AI-based capabilities get rolled out and trigger subsequent changes.

• Purpose: Align AI initiatives with the organization’s values and goals, connected to a clear sense of purpose.

• Ethical: Uphold strong ethical standards in the deployment and use of AI.

• Boundary-Spanning: Work across different departments and stakeholders to ensure cohesive AI strategies. In summary, our research across our membership shows that they are optimistic but clear-eyed about AI. They feel a sense of urgency in developing the proper guardrails and effective governance in order to set employees free to experiment and innovate. They emphasize education and communication so leaders can make informed investment decisions aligning AI strategy with their organization‘s purpose. But even the skeptics are unwilling to risk sitting on the sidelines. This technology is evolving at an unprecedented pace, and those who fall too far behind are unlikely to experience a grace period to catch up. There is a clear preference to lead by building trust, expressing empathy, and aiming for positive change.