Global Digital Foundation
The Global Digital Foundation is a think tank that offers policymakers a framework for dialogue with stakeholders and with their counterparts in other countries so that they can develop a shared understanding of the challenges of policy in a digital world.
ID: 352788324603-77
Lobbying Activity
Response to Requirements for Artificial Intelligence
24 Jul 2020
Policy analysts and policymakers are responding to AI as both a threat to human rights and as a potential savour of humanity from discrimination. This response rests on two questionable ideas. The first is that AI’s supposed human-like intelligence could give it a mind or a ‘will’. The second is that decisions affecting socio-economic outcomes are, in reality, the instrumentalisation of unfair hierarchies—and that AI could make this problem worse. This paper argues that, while AI affecting safety—e.g. in transport or healthcare—should be regulated and certified prior to implementation, AI that may affect human rights or social outcomes cannot be regulated in this way. This is because the likely causes of bad human rights or socio-economic outcomes following the use of AI will be both multivariate and exogenous to AI technologies in a way that threats to safety will not. Proposals to regulate AI for ‘fairness’ ex ante rest on a univariate ‘social bias’ explanation for differential social outcomes and are, hence, likely to be both ineffective and divisive.
Therefore Option 1 a 'soft law' approach is the best way forward.
Read full responseResponse to Fairness in platform-to-business relations
20 Nov 2017
The European Commission has identified a number of undesirable practices which it says threaten Europe’s successful transition to the Digital Single Market. While the Commission’s study and stakeholder consultation has found some questionable practices, fixing them will not require the more intrusive of the economic regulations which its “Reception Impact Assessment” is contemplating. There are three reasons for this. First, the practices, such as they are, don’t seem to be all that widespread—the evidence offered is anecdotal. Second some of the “problems” may not be problems at all. And third, mechanisms such as existing competition law and self/co-regulation offer a better alternative.
There are areas where platform businesses need to be more responsive to the needs and, contractual rights, of their businesses users. Unexplained delisting of products, or unnotified changes to Ts & Cs that have a material impact on businesses that are, thanks to platforms’ own success, dependent on their service, are indefensible. The Commission is right to take up the cause of eliminating such practices. But before we consider what needs to be done we should first ask two questions.
First, how bad is it? The Commission’s own survey reports that seventy six per cent of business users surveyed prefer to use redress mechanisms offered by the platform company rather than an outside mechanism. Eighty per cent of users feel that the Ts & Cs of platforms are fair, neither fair nor unfair, or have no opinion. The 20% who feel that the Ts & Cs are unfair suggests an operational problem rather than a systemic market failure requiring significant regulatory intervention. The report provides different figures for why ‘heavy’ and ‘non-heavy’ users of platforms feel the Ts & Cs are unfair. The dominant reasons for dissatisfaction of this 20% as a rounded % of all users is:
No possibility to negotiate or amend Ts & Cs (16% of heavy users, 12% of non-heavy users)
Possibility of one-sided changes by the platform (11% of heavy users, 7% of non-heavy users)
Limited access to dispute resolution (10% of heavy users, 5% of non-heavy users)
Unfair pricing (7% of heavy users, 6% of non-heavy users)
Biased or non-transparent search practices (8% of heavy users, 4% of non-heavy users)
These are not compelling figures.
The second (more provocative) question is: Is all of this really a problem at all? The Commission’s research has not quantified the economic effect of the problems that it has identified. The report tells us that the question about impact was ‘misunderstood’ by the survey recipients and so has not been included in the report. This leaves out the most important information of all. Further no cognisance is taken of the impact on consumers. These two facts alone undermine the case for sweeping regulation at the upper end of the Commission’s Impact Assessment menu.
There is no doubt that many of the businesses who responded to the Commission's survey have real grounds for complaint. But how much of the response from business users reflects general anxiety about the move to online trading that many of them have been forced to undertake? How many are simply complaining about an enforced change to their business model that is disruptive to them but not bad for their customers?
For example the lack of ability to negotiate or amend Ts & Cs is only a problem if those Ts & Cs are unfair in the first place. If the Ts & Cs don’t distort competition then there is no case for the Commission to act. Do standard Ts & Cs provide better outcomes for consumers? To what extent do the practices of the main internet platforms protect consumers?
There are, of course, very likely real potential problems behind the figures presented. The evidence suggests that existing Competition law and the possibiity of self/co-regulation are the best way forward.
Read full response