Dataiku’s tagline, Everyday AI, Extraordinary People, combines two complementary notions. The first, Everyday AI, expresses our vision for a future where organizations can use AI so effortlessly that it becomes almost pedestrian, powering their operations the way that electricity silently powers our lives without as much as an afterthought. The second notion, however, is even more important. The notion of Extraordinary People is fundamental to the vision that Dataiku has for this future of AI. It is not a future where humanity delegates decision making to cold machines, but where the people who use those machines do extraordinary things for themselves, their employers, and their communities.
This conviction comes from both our values and our observations. The fact of the matter is that we want to build a future that elevates humanity, while also ensuring our customers are able to build valuable AI throughout their operations. Far from contradictory, those two objectives are complementary for the following four reasons.
1. Internalizing AI Development and Use Is Becoming a Core Differentiator
Regardless of a company’s size, industry, or age, the ability to develop AI will become a core differentiator relative to their competition. As the use of AI becomes more generalized, the best marketing teams, the best logistics teams, the best engineering teams, the best finance teams, and the best HR teams will all be developing and using AI throughout their operations. This is because AI is a general purpose technology, with broad and deep applications anywhere there is data to be learned from.
McKinsey has argued that organizations can no longer hire and outsource their way out of the tech talent problem. The team that they have on staff today needs to become the AI building and deploying team of tomorrow. It requires effort, investment, and commitment, but it is necessary. The longer organizations delay building this capacity internally, the more time they are losing to the competitors who are investing in these skills today.
2. Employer Responsibility and Its Shareholder Benefits
Dataiku firmly believes that all organizations, ourselves included, need to work in the interest of both their shareholders and a broader range of stakeholders. Within that broader range of stakeholders, employees are an absolutely essential constituency. Given the power that employers wield over the lives of their employees, even in times when power may shift relatively in the opposite direction, there is a moral obligation to help employees grow and develop their skills.
AI has the potential to make work far more efficient and more interesting for humans. It will automate many of the mundane and rote tasks that we drudge through every day. It will allow us to focus our creativity and critical thinking on areas that require our unique skills while delegating to AI the parts that do not. That said, even OpenAI CEO Sam Altman acknowledges that certain roles will be eliminated in the future. Confronted with that reality, employers have a moral imperative to accompany their employees through this transition so that they can become the AI builders of the future. Finally, in addition to this ethical argument, upskilling existing employees also contributes to improved employee retention, bolstering the bottom line and delivering returns to shareholders.
3. Responsible AI Requires Upskilling
One of the tenets of Responsible AI is ensuring that there is a diverse range of voices who can participate in the development and implementation of the AI. We all have blind spots in our ability to assess whether AI is being developed properly — for example, if the training data has been appropriately collected and is unbiased — and if it is being used properly. We must involve a wider range of stakeholders to minimize the risk that the AI that we are deploying to improve our businesses is not inadvertently causing harm.
In order to ensure that this broader range of voices can have meaningful input on the process, it is important that they are familiar with how AI works. This does not mean that they need to become data scientists, rather, they need a high-level, conceptual understanding of the process. With this familiarity, these experts can provide better and more nuanced input into the development process, increasing the likelihood that the resulting AI will be unbiased and is not causing unintentional harm.
4. Better Business Results
Finally, we reach the bottom line. Upskilling its workforce allows a broader range of employees to participate in the process of creating and using AI. Not only is more AI being created and deployed, improving more areas of the business and therefore helping the bottom line, but that AI is better than if it would have been created only by an isolated team of experts.
This is because the quality of an AI system in an enterprise is not defined simply by some metric about the model, it is defined by the relevance of the results to the business. This is an important distinction and one that can be hard for some experts coming from academic backgrounds to grasp. The best model may not be the most sophisticated or the one with the best metrics, but the one that serves the business best. Identifying the one that serves the business best requires the expertise that can only be gained by working for years and even decades in a particular field. By upskilling these domain experts, they can help create that more relevant AI everyday, making them the extraordinary people of their organization.