A Call For Algorithm Bounty Hunters

Amid fears that the online environment is becoming dominated by machine-generated bias and deception, an article in the Canadian publication Motherboard proposes a market-based solution based on a version of so-called white-hat hackers. In this scenario, a similar group of “algorithmic auditors” would effectively become “the immune system for the age of algorithmic decision-making.” This is a necessary step, according to researcher Amit Elazari Bar On. She maintains that regulation and legal action will not be up to the task. One problem is that under current law those who do the research required to decode the potentially offending algorithms risk legal liability. What’s needed is to establish market incentives, including an “algorithmic bug bounty,” in order to develop a kind of marketplace that will foster algorithmic auditing, as we now have for research into security vulnerabilities. The writer is a doctoral candidate at UC Berkeley School of Law and a grant recipient from the Center for Long-Term Cybersecurity at Berkeley School of Information.