Hacker News Reader

Hacker News

Comment by golol | Hacker News Reader

Parent Full thread

golol

People always only think about the downsides of AI deployment and completely neglect the upsides.

> Callers to Portugal's 112 emergency line sometimes wait five to six minutes to get a response.

Improving this would be great!

> If the trial periods goes well, the AI will come into wider use by 2025, a senior politician said.

If issues are much more frequent than with human responders, the system will obviously not be deployed.

View on HN

These types of decisions come to be from only considering the upsides. Anyone who considered a single downside would have reasonably concluded that this is a terribly unwise decision.

> If issues are much more frequent than with human responders, the system will obviously not be deployed.

Part of the issue is that they're using metrics from a poorly supported system as comparison. If they funded the emergency service optimally then these comparisons would make sense.

They could easily reduce those response times through increased funding, training, and recruitment though... this is just a bandaid on a fundamentally dysfunctional pubic service in need of more state support

Most problems can be solved with increased funding, training and recruitment. But there are just not enough people. Yes, we could surely employ enough emergency call reaponders, but in many many governmental institutions you find a lack of employees, and our society can simply not afford to employ enough bureaucrats, teachers and emergency call reaponders. These issues will simply never be resolved via political means because they are an instance of plain economic scarcity. You can always say that X can just be fixed with more funding, but you can not do that for every single one of these subjects without balooning the government expenses 3x. Technological advances have a chance to address the core scarcity and therefore actually solve the issue. If we can double the total labor of people employed by the governemnt in such roles, then we have just created a massive gain for society.

I'm not sure why the level of analysis is appropriate for the context here. We have an extremely specific objectives with metrics (call wait times and calls correctly resolved). There's nothing fundamentally scarce about the labor involved in responding to 112 calls, it seems like a job pretty poorly suited to automation, we as a group decide how to allocate our resources, and if we want to cut tax rates at the expense of ambulances getting to people with heart attacks, that's the right of the Portuguese population; but let's not pretend that this is any kind of revolution in business theory

Reply on news.ycombinator.com