“We are not against the deployment of new technologies in a sensible manner, but we would oppose the use of AI if the primary motivation is to secure savings only,” said Simola. “Whatever the government decides on AI in the public sector transformation, the final implementation plan should not overlook what impact the changes will have on employees and on how public services are both managed and delivered to ensure quality and higher productivity.”
If AI can do the same job with equal quality, why would it be bad if the primary motivation is to secure savings only?
I’m not Finnish but as a citizen i would rather have my tax money spent on stuff that can’t be automated then on stuff that can.
Because if your prime factor to optimize for is cost savings “AI can do the same job equally well but cheaper” quickly becomes “AI can so a somewhat worse job but much cheaper”, then becomes “AI does quite a shitty job, but that’s what we have to live with now as we laid off all workers to save money”…
(Also thanks to capitalism it then continues with “Now that AI is the only option in that sector, let’s raise prices to increase our profit margin”.)
its a government so they don’t have a profit margin?
As long as the government pays public money on jobs in the public sector there isn’t. The moment that government tries to replace workers with AI –that they definitely not build, train, probalby not even run themselves– there is a profit margin, for the AI companies. Also now for the first time the public money spend to provide public services is no longer going back the tax payers in the form of wages but flow out to private companies. And it’s about AI so out of country, too.
As with everything, same with governments shoveling millions if not billions into microslop products it should be open. Open source/weight AI is very capable nowadays so it shouldn’t be an issue.
I don’t expect the government to be able to go into fully training their own foundational model, but a few racks in a government building somewhere can run open source/weight AI for a fraction of the cost. And is fully sovereign too instead of sending it to some provider.
Public money shouldn’t go to private corporations.
Governments spend less = other services get funded more, national deficit gets lower, taxes get reduced. Depending on what the government decides to spend the savings on.
Sure, if the starting position was 0 but that’s not the case. They are trying to cut spending because they are making huge losses. The financial situation in Finland is tough right now, really tough.
Because the alternative to saving money with AI is providing better services with AI - if it can be done. And the people to do that are the people doing the job.
If a human clerk handling applications or something makes an error 1 in a 100 applications and an AI system has the same error rate but is far cheaper. Then quality stays the same.
Because now you exluded the option of having the employee choose which applications to have Ai process and which need human review - the option that delivers the most quality
No? Where have I excluded that option?
If for normal volumes 100 employees are need to process. And with AI only 50 are needed, then it’s still cost saving?
I’m making up numbers because I don’t have the actual numbers and 2x is easy for the example, insert whatever numbers you like.
The issue with that statement is that putting double the employees on a project doesn’t always improve its quality.
If we go with the numbers we have been going with and a employee with AI is twice as productive then adding AI effectively doubles the time available. This only leads to a quality increase if work was being cut short. If each employee already got the required time necessary to handle a application with full quality then you haven’t increased quality but the capacity. If the now increased capacity is far above normal volume it is a waste to have so much excess capacity.
If we accept that employees spend the same amount of time as without AI but now the error rate is reduced then yes. I don’t think that is the case but I don’t have the numbers so who knows.
It isn’t inconceivable that they have picked up on the “hype-wagon” that AI is being propelled with. Investors have poured a lot into the technology and consistently promote the technology while all other sectors of society have mixed feelings about it. Still I agree that I can’t see why a financial incentive is being ruled out. Maybe it is based on the idea of putting people first, and not selfishly cutting jobs from society, then again, that would invalidate the idea in the first instance.
If AI can do the same job with equal quality, why would it be bad if the primary motivation is to secure savings only?
I’m not Finnish but as a citizen i would rather have my tax money spent on stuff that can’t be automated then on stuff that can.
Because if your prime factor to optimize for is cost savings “AI can do the same job equally well but cheaper” quickly becomes “AI can so a somewhat worse job but much cheaper”, then becomes “AI does quite a shitty job, but that’s what we have to live with now as we laid off all workers to save money”…
(Also thanks to capitalism it then continues with “Now that AI is the only option in that sector, let’s raise prices to increase our profit margin”.)
That’s why I emphasized if, there are plenty ways to get lower quality with human employees aswell.
I don’t really understand the last paragraph, its a government so they don’t have a profit margin?
As long as the government pays public money on jobs in the public sector there isn’t. The moment that government tries to replace workers with AI –that they definitely not build, train, probalby not even run themselves– there is a profit margin, for the AI companies. Also now for the first time the public money spend to provide public services is no longer going back the tax payers in the form of wages but flow out to private companies. And it’s about AI so out of country, too.
I said the same in another comment already.
As with everything, same with governments shoveling millions if not billions into microslop products it should be open. Open source/weight AI is very capable nowadays so it shouldn’t be an issue.
I don’t expect the government to be able to go into fully training their own foundational model, but a few racks in a government building somewhere can run open source/weight AI for a fraction of the cost. And is fully sovereign too instead of sending it to some provider.
Public money shouldn’t go to private corporations.
Profit cannot come before people. Otherwise Finland becomes America.
I agree, but the article is about the Finnish government, unless there is something I don’t know they weren’t making any profit to begin with.
They’re just trying to spend less, which is functionally the same.
Assuming the income stays the same
Corporations spend less = shareholders get richer
Governments spend less = other services get funded more, national deficit gets lower, taxes get reduced. Depending on what the government decides to spend the savings on.
How is it functionally the same?
Sure, if the starting position was 0 but that’s not the case. They are trying to cut spending because they are making huge losses. The financial situation in Finland is tough right now, really tough.
The profit here is for the AI provider supplying the service to the government.
No idea if the fins are based enough for this but i would hope they run open source AI.
Public money shouldn’t go to private corporations.
If only our system was not currently being demolished in the name of optimization.
Because the alternative to saving money with AI is providing better services with AI - if it can be done. And the people to do that are the people doing the job.
Why are those mutually exclusive?
If a human clerk handling applications or something makes an error 1 in a 100 applications and an AI system has the same error rate but is far cheaper. Then quality stays the same.
Because now you exluded the option of having the employee choose which applications to have Ai process and which need human review - the option that delivers the most quality
No? Where have I excluded that option? If for normal volumes 100 employees are need to process. And with AI only 50 are needed, then it’s still cost saving?
If we’re making up numbers my point is that with 100 employees using AI you get 2x the quality than 50 employees using AI
I’m making up numbers because I don’t have the actual numbers and 2x is easy for the example, insert whatever numbers you like.
The issue with that statement is that putting double the employees on a project doesn’t always improve its quality.
If we go with the numbers we have been going with and a employee with AI is twice as productive then adding AI effectively doubles the time available. This only leads to a quality increase if work was being cut short. If each employee already got the required time necessary to handle a application with full quality then you haven’t increased quality but the capacity. If the now increased capacity is far above normal volume it is a waste to have so much excess capacity.
Hardly, you need to maintain the same numbers to double check the work of AI. Haven’t improved throughput, but this guarantees better quality
If we accept that employees spend the same amount of time as without AI but now the error rate is reduced then yes. I don’t think that is the case but I don’t have the numbers so who knows.
It isn’t inconceivable that they have picked up on the “hype-wagon” that AI is being propelled with. Investors have poured a lot into the technology and consistently promote the technology while all other sectors of society have mixed feelings about it. Still I agree that I can’t see why a financial incentive is being ruled out. Maybe it is based on the idea of putting people first, and not selfishly cutting jobs from society, then again, that would invalidate the idea in the first instance.