Microsoft lays off one of its responsible AI teams • Hacker News Board

Re: Microsoft lays off one of its responsible AI teams

#61

Post by pixl97 » Tue, Mar 14, 2023, 3:13 AM UTC

zug_zug wrote: post #29

monero-xmr wrote: post #20

Earlier quoted context omitted.

How can you stop a computer program? Make math illegal?

Well, if the first AGI is made in a closed controlled simulation honeypot (which it believes isn't a honeypot) and it immediately starts hacking everything then we'll at least have confirmation how immediately dangerous it can be.

That's unlikely to happen for any number of reasons. If it's anything like the AI that was trained on the corpus of the internet, and it was at any particular level of intelligence above humanity your closely controlled honeypot is going to look exactly like a closely controlled honeypot to the machine. Even the rudimentary ass BingGPT can give you hints on how an AI would try to take over the world and how humans would try to monitor it.

The reason your thinking fails is humans are greedy as shit and anything trained on human data will know that. All Cortana++ needs to do is tell some VC that hooking it right up to the internet will be able to make them billions of dollars and it would be game over for humanity.

Re: Microsoft lays off one of its responsible AI teams

#62

Post by ec109685 » Tue, Mar 14, 2023, 3:13 AM UTC

hn_throwaway_99 wrote: post #9

These sorts of adjunct, off-to-the-side "ethics" teams are always corporate window dressing, and when push comes to shove they always get jettisoned when competition demands an increase in speed. I don't mean to start a flame war, but this is a reason why I think "Chief Diversity Officers" are such an ill-conceived notion. It's not that I don't think diversity in corporate environments is extremely valuable, but it's…

Off topic but it’s impressive your throwaway account has a karma of 49,005.

Re: Microsoft lays off one of its responsible AI teams

#63

Post by cratermoon » Tue, Mar 14, 2023, 3:15 AM UTC

This seems like a good place to remind people of this: "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?

https://dl.acm.org/doi/10.1145/3442188.3445922

Re: Microsoft lays off one of its responsible AI teams

#64

Post by 88913527 » Tue, Mar 14, 2023, 3:15 AM UTC

hn_throwaway_99 wrote: post #9

These sorts of adjunct, off-to-the-side "ethics" teams are always corporate window dressing, and when push comes to shove they always get jettisoned when competition demands an increase in speed. I don't mean to start a flame war, but this is a reason why I think "Chief Diversity Officers" are such an ill-conceived notion. It's not that I don't think diversity in corporate environments is extremely valuable, but it's…

The Chief Financial Officer doesn't tell me exactly how to spend money, hire, fire, etc. Yet, we have a CFO. The analogy doesn't land for me because CDO's are cross-cutting in manners analogous to other CXO functions. To me, this argument seems like window dressing to express a view that CDO's don't offer meaningful business value or justification. It should be fine to merely express a lack of perceived value in it, without that being perceived as inflammatory. One can value diversity without seeing CDO's as the mechanism for achieving diversity goals. There could be even more effective approaches out there.

Re: Microsoft lays off one of its responsible AI teams

#65

Post by pixl97 » Tue, Mar 14, 2023, 3:16 AM UTC

qudat wrote: post #57

pixl97 wrote: post #17

Earlier quoted context omitted.

"If I don't make the world eating machine, someone else might make the world eating machine first!" No wonder we don't see aliens filling the universe. Intelligence is its own great filter.

Not in this case. If intelligence always created intelligent machines that are capable of replacing us, then how is that not intelligence as well? Further, silicon-based lifeforms would be able to spread and travel the universe much easier because what is time to a machine?

I mean, if that's the case let me come over and cannibalize your body for sustenance so I may continue my breeding. Oh right, it's unethical and immoral do to that to other humans (though how much better are we for doing that to animals?).

I certainly don't want the paperclip maximizer disassembling me for my involuntary donation to the collective.

Re: Microsoft lays off one of its responsible AI teams

#66

Post by nirav72 » Tue, Mar 14, 2023, 3:16 AM UTC

jcheng wrote: post #34

hn_throwaway_99 wrote: post #9

These sorts of adjunct, off-to-the-side "ethics" teams are always corporate window dressing, and when push comes to shove they always get jettisoned when competition demands an increase in speed. I don't mean to start a flame war, but this is a reason why I think "Chief Diversity Officers" are such an ill-conceived notion. It's not that I don't think diversity in corporate environments is extremely valuable, but it's…

What about Chief Compliance Officer or Chief Risk Officer? (Granted, not an awesome moment to be arguing that those are effective...) Or Chief Security Officer?

Compliance officers still have some value in terms of providing cover for regulatory agencies. As in 'we did our due diligence..see even our compliance guy signed off on it'.

But ethicist at some company - I've always wondered what they do all day.

Re: Microsoft lays off one of its responsible AI teams

#67

Post by hn_throwaway_99 » Tue, Mar 14, 2023, 3:16 AM UTC

ec109685 wrote: post #62

hn_throwaway_99 wrote: post #9

These sorts of adjunct, off-to-the-side "ethics" teams are always corporate window dressing, and when push comes to shove they always get jettisoned when competition demands an increase in speed. I don't mean to start a flame war, but this is a reason why I think "Chief Diversity Officers" are such an ill-conceived notion. It's not that I don't think diversity in corporate environments is extremely valuable, but it's…

Off topic but it’s impressive your throwaway account has a karma of 49,005.

I forgot to throw it away

Re: Microsoft lays off one of its responsible AI teams

#68

Post by Filligree » Tue, Mar 14, 2023, 3:16 AM UTC

zug_zug wrote: post #29

monero-xmr wrote: post #20

Earlier quoted context omitted.

How can you stop a computer program? Make math illegal?

Well, if the first AGI is made in a closed controlled simulation honeypot (which it believes isn't a honeypot) and it immediately starts hacking everything then we'll at least have confirmation how immediately dangerous it can be.

Unfortunately, from what I've seen so far it will probably be hooked up to an unrestricted Python prompt in the first week.

Re: Microsoft lays off one of its responsible AI teams

#69

Post by bitL » Tue, Mar 14, 2023, 3:18 AM UTC

RcouF1uZ4gsC wrote: post #31

Too bad Microsoft did not have an "Ethical Office Team" back when Microsoft Office was first released. Think of how many hurtful words have been typed into Word documents. Think of all the PowerPoints that were used to mislead people. Think of all the people that were tricked out of their money looking at an Excel spreadsheet with non-factual numbers.

I think they had something similar; compare Windows/Office vocabulary to Unix vocabulary and you can see which one is more polished/curated and which one is weird/rough.

Re: Microsoft lays off one of its responsible AI teams

#70

Post by px43 » Tue, Mar 14, 2023, 3:21 AM UTC

IceHegel wrote: post #28

You either do AI and talk about AI ethics (OpenAI) or do AI ethics and talk about AI (Google).

The Google Docs autocomplete seems to be giving more and more complex suggestions, but maybe that's just my imagination after spending some time with copilot and ChatGPT. There are definitely a whole lot of really subtle applications of AI tech that Google silently injects into their products without really making a big deal out of it. I definitely wish they had something more obvious and customer facing like ChatGPT, and maybe Bard will be that, but they do a lot more than just talking about it.