Dr. Federica Carugati wrote this fascinating piece for Wired, suggesting that the key to making an AI-driven future equitable is to take some ideas from the ancient Athenians.
Machine-learning algorithms cover everything from online dating to media consumption and the criminal justice system, but racial and socioeconomic discrimination are major risks that come with it. There exists no true regulation for these algorithms; the closest thing to regulation are “journalists and civil society organizations” that “report when things have gone wrong.”
As the use of these algorithms grows more common, the need for regulation, in turn, also becomes more pressing. OpenAI, an AI research and development company, “suggests creating external auditing bodies to evaluate the societal impact of algorithm-based decisions.” Given that public policy relies more on real people making trade-offs for the groups they represent, it is harder to allow AI to follow just a series of logical rules to weigh in on it. For that reason, giving citizens a say on the matter acts as a solution to the problem of letting AI make trade-offs on its own.
Recent social experiments have shown that diverse groups of citizens are capable of making decisions about complex issues, and a historical example of citizens making decisions for the larger group can be found in ancient Athens, “an entire society [built] on the principle of citizen governance.” The Athenian model allowed for “a broad section of the population” to be represented and for the “diverse knowledge” of its members to be shared. Similar to juries, the citizen council could “act as an authoritative body or as an advisory board to an existing regulatory agency” and assess many issues including privacy, safety, security, ethics, data/labor sources, and more. Having a citizen council could “expand the range of possible solutions to the problems that algorithms create, enhance democratic accountability, and foster citizen participation and learning.”
Image Source: Wired Staff: Getty Images