Just two months after its launch in November 2022, ChatGPT had attracted more than 100 million active users, outpacing recent innovations and helping democratize generative artificial intelligence technology.
Large technology companies as well as businesses and start-ups across industries are rushing to carve their share of the transformational opportunity. While there is broad consensus on generative AI’s potential, there is also a shared sense of duty to develop and deploy it responsibly.
The latest capabilities powering generative models enable the creation of unique or contextualized content—including text, art, music and even entire virtual worlds—based on users’ prompts. The ability to generate human-like outputs that require a deeper understanding and creative thinking strengthens the disruptive potential of generative AI.
Generative capabilities will broaden and deepen the universe of applications compared to previous generations of AI. Many of these use cases have the potential to completely change how we work and create huge productivity gains by bringing automation to tasks previously thought to require human involvement. In Canada alone, generative AI has the potential to significantly impact the day-to-day activities of roughly 40% of the workforce.
Additionally, the pre-conditions are in place for generative AI to scale quickly and diffuse across the workforce. With strong penetration of the cloud and easy access to low-cost computing power, the technology infrastructure can be dialed up to support increased use. Moreover, generative AI’s built-in reinforcement effects will accelerate test-and-learn cycles, making more use cases available more quickly. The multiplication of use cases combined with strong perceived value from using the technology will foster human adoption. The transformation will be broad and deep and will happen fast.
Key ways generative AI is different from other enterprise technologies
Generative AI is unlike other technologies that organizations have had to integrate. It brings specific challenges and opportunities that will require leaders to act and mobilize their teams differently while focusing on value, agility and responsibility:
Opportunities to pursue use cases will seem limitless. This creates the risk that organizations chase the wrong use cases or spread themselves too thin—delaying the path to value from the highest-impact opportunities.
Organizations will need to get used to a world where no single technology solution fits all requirements. Indeed, different use cases will require different generative models—and the supporting technology and data capabilities to develop and deploy them.
Regardless of the use cases pursued, the workforce impact is likely to be broad and deep. New and vast pools of workers will need to adapt. Many will be skilled workers (e.g. software engineers, legal clerks, financial advisors) in jobs AI couldn’t replace in the past.
By developing customized language models and integrating proprietary data into them, innovative organizations will generate new intellectual property assets that will need to be proactively protected, managed and, preferably, commercialized or monetized to deliver business value.
Leveraging generative AI capabilities at scale requires new foundations, such as a modern data architecture that can integrate data from different sources, a cloud-based infrastructure that is flexible and efficient and new skill sets to translate use cases into reality.
Generative models are broadening the risk landscape, requiring additional controls on both model inputs and outputs as well as strengthened cybersecurity, data protection and privacy protocols. Risk executives will need to strengthen and embed agility in their lines of defence in order to adapt to an evolving risk landscape.
Siloed experiments focused on specific use cases will no longer be sufficient to harness the true potential of generative AI. The ability of foundation models to enable multiple use cases across the organization demands a more coordinated approach among cross-functional teams to explore the possibilities and test and learn together.
To create value over the long term, leaders should adopt a holistic approach
We believe winning organizations will adopt a holistic approach based on four key pillars.
The first pillar consists of anchoring the organization on a clear path to value through generative AI, using a value compass to guide use case selection
Investments should focus on a prioritized portfolio of applications designed to both create productivity gains and drive competitive advantage through differentiating use cases based on proprietary data. Leaders will need to proactively assess and prepare for the workforce implications of adopting generative AI to unlock value. This includes adjusting workforce planning to account for new skill sets that are needed as well as potential redundancies, investing in upskilling and reskilling programs to support generative AI integration and creating a culture that rewards experimentation and continuous learning.
The second pillar aims to enable the development and use of generative AI based on trust and responsibility.
Leaders should define and socialize guiding principles from the onset to embed responsibility in all development and deployment activities and prevent misuse. For instance, principles can cover access rights, approval processes for model iterations and restrictions on the use of certain types of data. Those guardrails can then be used as a framework to augment risk controls for cybersecurity, data protection and privacy, create processes to supervise model inputs and outputs and check for bias and explainability.
The third pillar focuses on future-proofing technology and data capabilities so that they can effectively power the transformation.
Organizations will need to develop or acquire capabilities, such as prompt engineering and fine-tuning, to deliver against specific requirements of generative models. Technology leaders will also need to enhance their technology and data environment with the right infrastructure and sufficient resources (e.g. computing power) to handle a portfolio of models, each tailored to meet the specific requirements of different use cases. Organizations already operating in a multicloud environment may seek to harness the capabilities of multiple providers to meet their requirements optimally.
The fourth pillar involves working with an ecosystem of (in some cases multiple) partners to enable their transformation.
They should define where their capabilities are best fit to support their ambitions and map out areas where partnering with providers, both hyperscalers and specialized solution providers, would be more effective or resource efficient.
What actions can leaders take to kickstart their journey?
Unleashing the full potential of generative AI will take time, but there’s an urgency for organizations to act. So how can you kickstart your journey? Looking back on key lessons learned from the first eight months of market activity and from our own journey at PwC Canada, we’ve developed a perspective on how leaders should define their organization’s approach to generative AI. Download the full article to learn more and explore seven key actions to help organizations get started.
Download the full article
to learn more and explore seven key actions to help organizations get started
A community of solvers to help guide your generative AI journey
To help Canadian leaders start the journey, our community of solvers is sharing their insights into the key issues raised by generative AI, including the very important considerations related to trust, ethics, privacy and consent. They’re also helping organizations explore the types of use cases at play and the benefits they create, the very significant implications for the workforce and the technology capabilities and partnerships required to move forward. We invite you to read our perspectives on the many questions surrounding generative AI and look forward to discussing the path forward for your business as you begin your own journey.
