3 ms·
Maybe I'll look like an idiot for saying these things in a few years. Maybe I'm naive and maybe I'm even a little bit cynical, but shouldn't we wait until there
by devjab 4y ago
Maybe I'll look like an idiot for saying these things in a few years. Maybe I'm naive and maybe I'm even a little bit cynical, but shouldn't we wait until there is something to have a debate about? From my anecdotal perspective we haven't seen a lot of areas where the recent AI "developments" are going to have a meaningful impact on how we do business. I work in a non-tech focused enterprise organisation which develops and sells solar, which involves a lot of data. Both in terms of running power plants, but also financial and legislative data. Much like the rest of the world, we've tried hard to implement GPT into our workflows, and we've frankly had very little success making it do anything useful. Ok, that is not entirely true, because it has been fairly successful at competing with our RPA software. Long story short, a lot of our software and a lot of the workflows in our company involve moving and manipulating data between systems. With GPT some of our more technically inclined workers have been able to successfully automate parts of their workflows in a way that no no-code solution has been able to do before. So there is certainly some success, but we were already doing these things with RPA and other no-code solutions, and the long term strategy for these things is still to be better at buying software that comes with API's so that we do not need "digital gaffatape". This is obviously never going to happen, and it is great that the "digital gaffatape" has improved, but it hasn't exactly blown our minds. But it has been an improvement.
That is, however, the end of the usefulness. In every other attempt, we've come up short, and it's not for lack of trying.
I work as an external examiner for CS students, however, and this is an area where I have seen a massive disruption. Because ChatGPT is really great at cheating. Not so great for the students who enter a final, and find themselves unable to answer things they would have, if they had written their work themselves, but for a lot of education I think there will be a need to rethink how we teach and test students. Which is a great thing, mind you, but how much money can that really be worth for these AI companies?
Anyway, I'm probably wrong, but I still think it's healthy to wait and see what happens before we waste too many legislative resources on it. Imagine if we had spent resources on the Blockchain hype. I'm sure a lot of people will line up to tell me how wrong I am about this comparison, but we regulate Blockchain quite heavily in the EU now. So it's not like the legislative branches can't grind into gear once something actually happens.
- dangond 4y ago> shouldn't we wait until there is something to have a debate about? The main push for regulation (at least from those worried about existential risk) comes from the fact that in the scenario that AI does get powerful enough to take over the world, it will be too late/useless to have a debate at that point. The only time where it makes sense to have a debate, from that perspective, is before the capabilities are developed. If such capabilities are developed, and we have no idea how to control such a system because we decided not to worry about it, humanity loses.
- devjab 4y agoIt is? Maybe I live in a bubble of some sort, but the only worry I see about AI is how automating too many jobs too quickly will impact society. Personally I'm not worried about the Terminator scenario, both because I don't see AI going in that direction at all, but mainly because we won't be able to regulate what nation states do with AI in terms of defense. That may be a little pessimistic, but we can't even regulate what they do with Nuclear Weapons or who they go to war with, so we'll never be able to regulate what they do with AI.
- goatlover 4y ago> Personally I'm not worried about the Terminator scenario, both because I don't see AI going in that direction at all, Not so much the terminator scenario, which was more a plot device to tell a time travel story with killer robots. But that a powerful enough AI(s) might do something unanticipated which is very harmful, and we would have little ability to control it. For now, it's more about how people might abuse AI or how it might disrupt society in ways we haven't predicted. Similar to the negative effects of social media. If the internet gets flooded with fake news that we have a hard time telling apart from the real thing, then it becomes a big problem.
- concordDance 4y agoWe aren't going to get superintelligent AIs immediately, but we do need to think about how to avoid them coming about as soon as possible so we have time to put our plans in to motion.