Artificial Intelligence technologies are ever more present in our daily lives; they are already impacting how we do commerce, both in the act of finding and purchasing goods and in creating smarter ways to advertise and sell a product online. Even the smallest businesses can enjoy these latest innovations thanks to online platforms that help bring them within their reach, like eBay.
Still, we are only at the start of AI’s story and policy-makers, civil society, users and companies need to drive an open dialogue to ensure AI progresses in ways that benefit humanity as a whole. As a leader in commerce and technology, and the largest global commerce platform for independent small retailers and entrepreneurs, eBay takes an active interest in this debate and shares the ambition of an artificial intelligence that:
1. Has a positive impact on the economy, for example by spreading wealth among businesses no matter their size or their location.
AI may disrupt the forms of work as we know them, but should always end up fostering healthy competition and innovation, so the net result for society remains positive.
In that sense, the increasing use of AI-based algorithms may generate concerns around unfair coordination between market competitors, for instance around price-fixing. Companies should remain responsible for antitrust law violations no matter if the means of coordination they are using relies on AI.
2. Is inclusive of all people, their cultures, ethics, social preferences and opinions.
AI should not be reserved to a privileged elite, nor should its decisions lead to manipulation or discrimination of a given group of users. AI could especially be used to benefit disadvantaged groups in wider social good or philanthropic efforts.
At eBay, inclusion is a core aspect of the company’s purpose and we are already helping independent small and micro businesses, including from remote or less advantaged regions, to participate in the global market. We hope to use AI to improve our efficiency in achieving that same goal.
3. Functions in a fair, accountable and transparent way.
Where AI will increasingly be used to supplement or replace human decision-making, this artificial process needs to remain rationally explainable to human beings. Not only will this make AI more acceptable to society, it will also allow developers to identify and fix potential biases that could otherwise be amplified through AI.
Transparency will also be an important tool in clarifying the legal liability regime, a central question to AI development. Robots and AI-enabled systems are tools that serve humans. When a damage is caused, and its origin can be attributed to the AI-enabled system, we believe transparency should be made on all potentially responsible parties. These will include the AI system’s user, designer or producer, and seller.
As we progress towards increasingly autonomous AI-enabled systems that react, learn and change largely independent from human programmers or directors, transparency will be crucial to understand the path these systems took to evolve and that led to the damageable actions in the case of legal challenges.
4. Brings as much security as human-led systems.
Policy-makers and companies alike will need to understand, manage and account for these new “hyper-data needs” in practices and regulation, all the while maintaining key goals like fairness, transparency and accountability.
5. Enables the greatest machine-human complementarity.
AI must focus on improving the natural capacities of people without replacing them, in other words to make problem-solving more efficient without taking away from it inherent and essential human perceptions.
In the world of commerce, such a debate has been known to take shape in price discrimination systems: the AI-reliant practice of companies to charge different prices to different consumers, based on some of their characteristics or behaviors.
eBay is an open marketplace that does not set the price for the items offered for sale. We believe AI finds better applications in areas where it does not directly try to replace free and educated consumer choice, for example when it looks at providing better recommendations or showing the full spectrum of value available on our platform.