It is common knowledge that big institutions lack the adaptability and speed to adopt to new changes, simply because of their sheer size and “proven” processes of how they conduct their business. Especially, governments are a prime example for this. Another key point for government is the involved bureaucratic workload that accompanies every decision and action, thereby slowing down its actions and decisions. So how could an implementation of AI help with that?
AI will allow citizens to have better digital interactions with public services, such as applications for crucial documents or other legal issues (Barroca, 2023). By which the human workforce will be exonerated, as the bureaucratic work are taken care by the AI (Barroca, 2023).
Another point of improvement would be the automation of routine tasks (Intel, n.d.). An example in Germany proves how big the impact could be. Students in Germany can apply for a student loan from the government (called BAföG). This application process was made entirely digital. However, these applications are then printed out by the personnel to be taken care of on paper instead of digitally (Funk, 2022). This results in an enormous delay in the processing and approval of such said applications, which could be minimized by the introduction of AI into the process (Barroca, 2023).
Eventually, AI will also enable politicians to make better informed decisions (Intel, n.d.), as its capabilities to process and analyse great amounts of data will allow policymakers to create better and more effective polices (Barroca, 2023).
To put it in a nutshell, AI could drastically improve the efficiency of government processes, as well as it would improve the effectiveness of imposed policies, thereby helping all involved stakeholders.
References
Barroca, J. (2023, May 10). AI And The Future Of Government. Forbes. https://www.forbes.com/sites/deloitte/2023/05/10/ai-and-the-future-of-government/
Funk, J. W. (2022, December 5). BAföG-Anträge: Digitalisierung mit fatalen Folgen. tagesschau.de. https://www.tagesschau.de/investigativ/funk/studenten-bafoeg-digitalisierung-buerokratie-101.html
Intel. (n.d.). The future of artificial intelligence (AI) in government – Intel. https://www.intel.com/content/www/us/en/government/artificial-intelligence.html
Hey Nick,
Interesting viewpoint. But, the answer to your question is ‘most likely not in the short-term’ in my opinion. Since the data government agencies work with are confidential most of the time, this data may not be shared with large institutions managing these models. And, the application of algorithms is heavily regulated in the Netherlands to prevent another” Toeslagen Affaire”. On the EU level, new AI regulation will ban the use of algorithms in high-risk scenarios often applicable to government agencies. Thus, widespread government use of AI may not be a reality till many years in the future.
Sounds interesting! Although, there are two points which may be highlighted here.
First, I would say top level governmental officials may not use the AI technology to optimize their workflow because their work should remain confidential. If they would go after generative AI, the privacy of data might be in danger and this may result, in worst case, in data leaks which would compromise entire governments.
Secondly, I would we able to agree that it would make sense to apply AI to administrative positions in an effort to offload administration employees. Governments, in order to reduce chances of invigilation from third parties, can invest in developing state AI algorithm, which then would be applied to local processes. However, the process implementation should be reviewed thoroughly to ensure there is no impact on data privacy. Also, there should be a strict protocol on how to use the said process and according measures to use when an outage is observed. Yet, while this may be a viable solution, I see it happening only in 10 to 15 years from now on.
Hi Nick & Mees,
First of all, thanks for sharing this insightful thought. As AI makes its way into all societal aspects, you would expect this trend to persevere on the governmental side. Though, my comment builds upon Mees’ answer to the question your title asks. I might be even a bit more pessimistic in terms of the time frame in which AI will create the almost perfect government. As nowadays governments (especially in the Netherlands) tend to be the most indecisive organs in a strongly divided country, the ironic part is they seem in a tremendous need of AI and its advantages for (at least faster and decisive) decision-making.
As governmental problems often require tailor-made solutions, like in the earthquake damage in Groningen, the Toeslagenaffaire and the Refugee crisis in Ter Apel, AI can (in my opinion) not outperform humans. The decisions made in these types of issues, often require empathy and emotions AI does not have (and will never get). So apart from what Mees has commented regarding the confidentiality concerns with civilian data, I personally do not see AI being implemented on a large-scale in complex governmental activities that are mostly concerning a country’s most difficult and wicked problems.