The shift of our economy and society online is taking place without referendum. What could go wrong? As it turns out, plenty. The data driven economy features powerful economies of scale, economies of scope, network externalities, and pervasive information asymmetries, each of which individually promises market failure; collectively, there is a perfect storm of factors pointing in that direction.

There is live streaming evidence of the reality of these theoretical concerns – rising corporate concentration, rising income disparities, and the presumptive unaccountability of the lords of tech as evidenced by behaviour such as Facebook CEO, Mark Zuckerberg, declining to accept the request of the British Parliament to appear before it. Dealing with the data-driven economy will require policy reforms across the board – policy should be technology neutral, so we need to go full horizontal. Some steps suggest themselves readily:

• Where Data/AI/ML is doing mechanical functions, ISO/private-sector/ expert-led standards setting is called for.

• Where Data/AI/ML is doing human cognitive/decision functions, competence regulation would be appropriate.

• Where Data/AI/ML has societal impact (e.g., surveillance), we will need to develop tripartite consensus framework with effective democratic oversight over the executive arm of government.

• Where Data/AI/ML has military applications, cyber-security in defense of sovereignty will become essential, together with the development of international conventions on non-use and proliferation of autonomous armed robots (for example).

• Where Data/AI/ML has major distributional impacts, including on the allocation of work between humans and machines, we will need to update, and in some cases rewrite, the economic policy framework.

• Where Data/AI/ML intersect with trade and investment, new approaches to international rules will be required in areas such as competition, strategic trade and investment policy, and the role of FDI in knowledge- based and data-driven sectors

Excerpted from Data and Digital Rights: More Questions Than Answers – But Enumerating the Questions is Essential