01 Most companies facing difficulties are actually not ready to use AI
CBR: One of the most critical decisions facing CEOs today is whether and how to use AI in their decision-making processes and across their entire organization. What are some important findings from IBM's research on this issue? Can you give some examples?
Dong Haijun: From the results of our "2023 Global CEO Survey", we found that although most (75%) of the surveyed CEOs believe that advanced generative AI will become a tool for business success, CEOs are engaged in intense intellectual struggles over whether to use AI in their own decision-making processes.
Specifically, some CEOs are already using AI in their decision-making processes. For example, 43% of surveyed CEOs said their companies are already using generative AI to inform strategic decisions. 36% said their companies are already using generative AI to support operational decisions. 50% said their companies are integrating generative AI into their products and services.
However, many CEOs also have very big concerns about AI, with data issues being the biggest worry. Even the most powerful AI can provide incorrect, biased or dangerous results without trustworthy and reliable data. 61% of surveyed CEOs are concerned about data lineage or data sources; 57% are worried about data security; 53% are concerned about data governance and compliance constraints; 48% are worried about data bias or data accuracy.
However, data issues are not unique to generative AI. Companies have been facing severe data challenges for years. According to research by the IBM Institute for Business Value, companies that rank higher in revenue, growth and technological maturity pay more attention to data standards and quality. Such excellent CEO decision-makers know that filling data gaps is a tedious and difficult task, but sound data remains an important priority for building competitive advantage.
CBR: Microsoft has proposed using artificial intelligence as a Copilot. For example, Nadella lets GPT help him write emails, draft memos, and even attend Teams meetings, which actually helps eliminate the boring parts of work. But at the decision-making level, AI may just be a good reasoning engine to assist CEOs in searching and thinking, and ultimately cannot replace the person making the decision.
Dong Haijun: Of course, that's why there's still a CEO position. Speaking of this, I also want to talk about the relationship between artificial intelligence and organizations. Currently, most companies are still in the "+AI" stage, not "AI+". "+AI" refers to using AI to replace human labor to complete some logical and relatively uncomplicated work. Soon companies will enter the next stage - "AI+", which is actually the reconstruction of organizational capabilities centered on AI.
In my personal opinion, the first area for companies to change from "+AI" to "AI+" is human resources.
To give a practical example, IBM's artificial intelligence is used very deeply in the field of human resources, and the task of automatically adjusting salaries for employees is judged and completed by artificial intelligence. It means that the traditional evaluation system will shift to an AI-centered judgment of employees. Artificial intelligence will give a recommended salary adjustment plan based on comprehensive factors such as the employee's market competitiveness, skills, future career development, past performance, and potential that can be developed. The release time of this plan will not be limited by the traditional salary adjustment cycle. It's like having an omniscient, all-powerful "brain" that knows employees very well, customizing salary and benefit structures, promotion and learning curves for employees.
CBR: While salary increases are certainly welcome by all, will using AI to assist in decision-making on the scale of layoffs and overall salary reductions become a scenario that many companies actually adopt?
Dong Haijun: I think most companies facing difficulties are actually not ready to use AI.
From an operational perspective, companies go through four stages: organization-driven, process-driven, data-driven, and finally AI-driven. These four stages cannot be skipped. The vast majority of domestic companies are in the first stage, which is the organization-driven stage. If they stay in the traditional organization-driven stage, that is, driven by functions and responsibilities, there will definitely be a lot of departmental walls and data inconsistencies, and the application of artificial intelligence will only be rejected by this organization. If a company's salary increase process is not clear and the data is not complete, how can the decision-making power be given to artificial intelligence?
CBR: Many CEOs face challenges in terms of workforce, corporate culture and governance in the process of implementing and expanding generative AI in their organizations. Are these challenges entirely new ones brought by new technology, or are they challenges that organizations originally faced but were "revealed" or amplified when new technology arrived?
Dong Haijun: It should be both.
The brand new challenges that generative AI brings to the workforce mainly include three aspects:
First, generative AI requires brand new talents and skills. Our survey found that currently, 51% of global surveyed CEOs and 49% of Greater China surveyed CEOs say they are recruiting for positions related to generative AI that did not exist in 2023.
Second, generative AI will create more positions. A large proportion (47% and 49%) of global and Greater China surveyed CEOs expect to reduce staff due to generative AI, but at the same time say that the number of jobs created will exceed the number of jobs lost. On average, both global and Greater China surveyed CEOs plan to increase their workforce by nearly 6% over the next three years.
Third, generative AI will create entirely new ways of working. To fully leverage the value of these new positions, talents, and skills, organizations must explore future ways of working. Organizations cannot integrate future talent into past operating models and must explore entirely new ways of division of labor and operating models.
Talent and capabilities have always been important challenges facing organizations, and the rapid development of generative AI has further amplified the shortage of new talent and skills. 53% of global surveyed CEOs and 56% of Greater China surveyed CEOs say they are already working to fill key technical positions, but the talent shortage is unlikely to ease in the short term. Both global and Greater China surveyed CEOs say that 35% of their organization's employees will need retraining and new skills training in the next three years, compared to only 6% in 2021.
For generative AI to move from improving productivity to innovating business models, technology is not the biggest problem. The biggest problem is that everyone in the organization truly uses generative AI, which is the challenge of corporate culture, mainly including three aspects:
First, the cognitive challenge. Many people think that generative AI is a tool that will replace their work, rather than a tool that can support their work and be used by them, so they have an inner resistance to new technology. If employees understand how this technology can make their work easier and more valuable, the adoption rate of generative AI may increase significantly.
Second, the challenge of adapting to and mastering new technologies. 64% of global surveyed CEOs and 66% of Greater China surveyed CEOs say their organizations must leverage technology that changes faster than employees can adapt, while 61% of global surveyed CEOs and 59% of Greater China surveyed CEOs say they are pushing their organizations to rapidly adopt generative AI, which makes some people uncomfortable. Most CEOs know that to fully leverage the effectiveness of generative AI, they need to develop technology and train people in equal proportions.
Third, the challenge of transformation. To foster a culture of accelerated transformation in the organization, 81% of global surveyed CEOs and 84% of Greater China surveyed CEOs say that inspiring teams to build consensus around a shared vision can achieve better results than providing precise standards and goals. However, 37% of global surveyed CEOs and 34% of Greater China surveyed CEOs admit that their organization's employees do not fully understand how strategic decisions will impact them.
The governance challenge refers to the need for organizations to establish comprehensive governance guardrails to ensure that employees can innovate within a safe framework as more and more people at various functional areas and levels of the organization use generative AI. 75% of global surveyed CEOs and 84% of Greater China surveyed CEOs say that achieving trustworthy AI is impossible without an effective AI governance framework in the organization, but only 39% of global surveyed CEOs and 38% of Greater China surveyed CEOs say their organization currently has a good generative AI governance framework in place.
02 CEOs are determined and optimistic, executives are hesitant
CBR: 70% of surveyed CEOs say AI has brought benefits to the entire organization, but only 29% of other surveyed executives believe their organization has the internal expertise needed to adopt AI. Which data is more revealing of the truth? Why is there such a huge "temperature difference" within the organization?
Dong Haijun: In fact, both of these data points are quite true. This huge "temperature difference" is caused by the "position difference", that is, standing in different positions within the organization, the perspectives and visions are different. Compared to other executives, CEOs stand at a high position in the organization, so their perspective is definitely wider and their vision is definitely farther. Other executives focus on their respective functional areas, so their perspective is more vertical and their vision is narrower.