The case for responsible AI design
Artificial intelligence has the potential to transform the way we work and live, by some estimates adding up to $13 trillion to global economic output by 2030—about 16% higher than it is today. But to responsibly harness that potential, and ensure its development is inclusive and does not further widen gaps in opportunity, policymakers need a complete understanding of AI’s effect on the global economy.
That’s why we’re excited to contribute our insights to the OECD’s AI Policy Observatory. The new initiative brings together a wide variety of stakeholders—including policymakers, technology companies, researchers, and NGOs—to create a forum to research and discuss how to develop AI responsibly.
Through this partnership, insights from LinkedIn’s Economic Graph are shedding light on how AI skills are reshaping the global workforce. Here are some of the trends we’ve found:
- The United States currently tops the world in AI hiring and talent development, but the rest of the world is catching up. In some cases, they’re hiring talent away from the U.S. Our research shows that in Brazil, Canada, and Australia, the average rate of hiring for people with AI skills has more than doubled in just two years, with hiring increasing 250-300%. Singapore and India have the next highest growth rates. To keep pace, the United States will need to make investments in attracting, developing, and reskilling talent at a large scale, or risk losing its leading position.
- In the U.S. and Singapore, startups and younger “digital native” workers are driving the supply and demand for AI talent. In Europe, individual industries are driving adoption of AI technologies. In countries where industry leads in AI, governments can concentrate on building hubs around these industries to help AI skills diffuse into other industries; and they can also work with business leaders and training providers to help ensure talent has the right skills to work in the jobs employers are hiring for.
- AI risks worsening inequalities if we don’t address gender gaps. Across almost every country, AI skills intensity for women is relatively lower than it is for men -- which means that women are less likely than men to work in occupations that currently use AI. This imbalance can create significant unintended consequences in systems that are more vulnerable to hidden biases in their developments.
But for all of the good that all of these new metrics and insights can do to help shape better policy decisions, we need more than data and insights to build an equitable future.
At LinkedIn, our approach to AI design aligns with the OECD’s Principles on AI. Under the umbrella of what we call “responsible design,” we work with outside experts and partners to avoid unintended consequences and build a platform that works for everyone.
This work spans all of the lines of business within our company, and includes cross-functional teams of designers, researchers, engineers, and product specialists. Understanding the real-world impacts of how data is collected, the AI systems that are trained on this data, and the products that are powered by AI is a hard challenge, but it is a moral imperative for all of us.
We’re still just at the very beginning of AI’s diffusion into every part of our work and lives, and leaders need more hard evidence and data to help steer the conversation and inform policy decisions. We’re proud to partner with the OECD to make evidence and data about the current state of AI -- which will help inform policy decisions that ensure all members of the global economy benefit from the changes that it will create in the new economy -- more accessible and more actionable.