AI and detection of competition / antitrust law violations
The UK Government has just issued a report, Growing the Artificial Intelligence industry in the UK. This blog comments on how the recommendations raised in the report can potentially aid competition enforcers in sharpening their detection tools in the AI realm.
A number of competition law scholars suggest that existing enforcement tools are not sufficient by themselves in terms of artificial intelligence (AI) because they cannot predict future markets or dynamics. The main practical challenges argued are:
- Competition enforcers are not aware of the problem (that is, they cannot detect AI collusion); or
- Competition enforcers are aware of the problem but lack the tools to fix it. For example, in behavioural discrimination, agencies do not understand new market dynamics and therefore cannot understand the shift to “almost perfect behaviour discrimination”.
The competing argument, however, is that these challenges are not new even in the AI realm and the proper tools exist but cannot be utilised in isolation. In publishing its recent report, Growing the Artificial Intelligence Industry in the UK, the UK Government offers eighteen recommendations that fall under the three following categories:
- Improving access to data;
- Increasing the supply of skills; and
- Maximising AI research.
Are these recommendations the potential solutions to the challenges competition enforcers face in terms of AI? Improving access to data can aid in the detection of anti-collusive behaviour or at least improve the detection of information exchange between competitors. Increasing the supply of skilled experts in AI can provide enforcers with a better-equipped task force. Finally, maximising AI research can sharpen regulators’ tools as they will understand more about what AI can do and where it can be applied. At the very least, it seems that these recommendations can sharpen competition enforcers’ current detection tools.
 Ariel Ezrachi and Maurice E. Stucke, 2016. Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy.
The author would like to thank Sandy Aziz, (Knowledge Assistant) for her assistance in preparing this legal update.