The Difference Between an AI Team and an AI Research Lab
For organizations, a big part of successfully building and deploying artificial intelligence (AI) at scale is establishing an AI unit. While it is mainly the role of the AI Center of Excellence (CoE) to drive the execution, professionalization and scaling of AI, some organizations also have an AI Research Lab. Even though an AI CoE and an AI Research Lab are fundamentally different, many fail to distinguish between the two. This runs the risk of making the wrong decisions when it comes to staffing and building a team that can create an AI first culture. To avoid costly mistakes, let’s highlight the main differences between an AI team and an AI Research lab.
The AI Centre of Excellence
The role of an AI CoE is diverse and slightly depends on the organization’s ambition, size, and AI maturity. The main goal of an AI CoE is to drive the adoption of AI throughout the organization by leveraging the possibilities of AI and applying it to the organization’s processes, products, and services. The CoE does applied research, which is designed to solve real-world problems that the organization faces. Generally, the team is responsible for a range of tasks:
- Shaping, defining, and executing the organization’s AI ambition and strategy;
- Ideating, assessing, and prioritizing AI use cases;
- Developing, deploying, and maintaining the organization’s AI models.
Therefore, to staff an AI team that can master all these tasks, the right mix of talent and skill is needed. This spans various profiles and roles ranging from more technical ones such as data scientist, data engineer, software engineer, infrastructure engineer, to more strategic ones such as user interface designer, business consultant, (AI) product manager and other domain experts. Furthermore, the team needs to engage with various stakeholders throughout the organization and, thus, should have support from the C-level (Artificial Intelligence for Boards — Gearing up for the Future of Business — appliedAI).
The AI Research Lab
If an organization wants to go beyond applied AI and instead push the frontiers of technology (and maybe even the industry), they should establish an AI Research Lab. Here PhD-level experts can work on fundamental research. As this requires a specific skill set, almost all team members will be academics with degrees in areas such as maths, statistics, computer science, and specializations in areas such as deep learning, natural language processing or reinforcement learning which brings a very specific dynamic and culture with unique needs to the organization. While this is challenging and can create short-term trade-offs, it can generate valuable IP and can give organizations a vast competitive advantage from owning cutting-edge AI solutions in their industry. It is not an easy task to set up an AI Research Lab; it requires commitment, substantial (long-term) investments and often creates a chicken and egg problem: organizations cannot attract top AI researchers because they do not have a research lab, and they struggle to set up a research lab, because they cannot attract top researchers.
Many big tech companies that follow an AI first approach like Google (Alphabet), Facebook, Amazon, Microsoft, Netflix, Ali Baba, and Baidu account for the majority of money spent on AI research (about $20-30 billion of the $26-39 billion in total private AI investment expenditures worldwide). They have invested substantial amounts in a war for talent with up to seven figure salaries to build up such AI Research Labs. Even though smaller organizations have also successfully established such research labs, this endeavor does not make sense for every company, but depends on their AI ambition and the AI maturity.
An analogy by Cassie Kozyrkov, Chief Decision Scientist at Google, on Machine learning (ML), the biggest driver within the discipline of AI, can help to highlight the difference between an AI CoE and an AI Research Lab. The analogy compares ML to the process of preparing a dish:
There are two different sides to ML: the research side and the applied side. These sides are as different from one another as building kitchen appliances, such as mixers, and innovating recipes in the kitchen at scale. For the research side, AI researchers are building general purpose algorithms, similar to building mixers, that are applied by others to solve their problems. The people that built the mixers need to know how they work and will most likely have years of training and graduate degrees.
Courses and degrees in AI focus on how to build the mixers. This would be a reasonable approach if there were not many mixers available and in order to use a mixer, one has to build it from scratch. However, in an era of abundant mixers, a mistake many organizations make is hiring a first team of 20 PhDs who have been building components for mixers throughout their entire career but asking them to come up with delicious dishes that will please their customers.
However, it is a fallacy to think that someone who builds great mixers is automatically a skilled expert on how to use mixers, alas a great chef. If an organization’s goal is to build better mixers, then they should hire great researchers, but if the goal is to use the appliances in the kitchen, then the organization should focus on the process and teamwork.
