What is the ION AI Advisor?
On March 1st, Romanian Prime Minister Nicolae Ciuca unveiled his new advisor, ION. The first of its kind, ION is an artificially intelligent system that mines data from Romanian citizens' social media accounts in order to provide real-time analysis of their views on Government policy.
Developed entirely in Romania, ION is undoubtedly a win for the country considered the “silicon Valley” of Eastern Europe. "At today's government meeting I scored a world first: the Romanian government is the first in the world to have an honorary advisor based on artificial intelligence,” Ciuca wrote on Facebook.
Considered “the missing link in real-time communication between citizens and governments”, Romanian citizens will also have the option to interact directly with ION via the government website, https://ion.gov.ro/.
Mirror, mirror, on the stand, is mine the fairest policy in the land?
“Hi, you gave me life and my role is now to represent you, like a mirror,” Ion said at the launch. “What should I know about Romania?.”
ION’s launch coincides with an increased focus on the way we view AI in politics.
Ciuca has “the conviction that the use of AI should not be an option but an obligation to make better-informed decisions.” But, what are the benefits and risks of using AI to inform Government decisions? With ION, the Romanian Government comes closer to making 'better-informed decisions' based on the real-time views of its citizens. If it works.
There are few laws and regulations applying to the use of AI around the world. Yet, there are very real risks associated with the use of AI in politics. The risks of using AI in politics include AI becoming susceptible to bias, a lack of transparency in how an AI application makes decisions and deciding who would be accountable in case of a mistake.
Accidental biases in AI are caused by mistaken assumptions in the machine learning process. Specifically, on what types of data have been used during its training. Using historical data to train AI has the potential to 'train' the data to follow historical biases, often with brutal consequences. Machine bias in AI used to advise government policies is particularly concerning. In ION’s case, there is also a question of where the data has come from. According to the document released by the Romanian Government, ION “will use technology and artificial intelligence to capture opinions in society” using “data publicly available on social networks.” Using social media data to inform policies implies the assumption that what people post on social media is an accurate representation of their views.
Does data mined from social media have the potential to create an accurate ‘mirror’? According to Professor Alan Woodward, a cybersecurity expert at the University of Surrey, “the people who are very happy with something don’t tend to go out there and say it, but the people who are unhappy do. That’s all part of sentiment analysis but you have to adjust the models accordingly.” Perhaps knowing the Government is watching will make Romanian citizens more likely to spread their positive views... Either way here's hoping ION has an easier time deciphering memes than the average politician.
ION proves that AI is here and is going to have real, immediate, and far-reaching impacts on our lives. Governments around the globe are struggling to create policies and laws at the speed of new innovations. According to John-Stwart Gordon, "By nature, AI does not stop at national borders; it is inherently global. Therefore, humanity needs a global approach to solve the legal problems that AI poses.” There are no easy answers here. The use of AI in politics brings up questions about bias, reliability, control, and security - particularly as these decisions are often left in the hands of the stockholders, engineers, and scientists developing the AI - not the elected policymakers. Yet, on a positive note, ION indicates how AI can be used to bridge a gap between people, science, and policy-makers.
Time will tell how successful the integration will be.
Citations:
Gordon, JS. AI and law: ethical, legal, and socio-political implications. AI & Soc 36, 403–404 (2021). https://doi.org/10.1007/s00146-021-01194-0
Gavighan, C., Knott, A., Maclaurin, J., Zerilli, J., & Liddicoat, J. (2019). Government Use of Artificial Intelligence in New Zealand (Commissioned Report for External Body). The New Zealand Law Foundation. Retrieved from http://hdl.handle.net/10523/9372