Live data from Hacker News

Google offers to help others with the tricky ethics of AI

arstechnica.com

31–35 of 35 posts

Re: Google offers to help others with the tricky ethics of AI

#31
post #30
post #27

Earlier quoted context omitted.

That example ( https://research.google.com/bigpicture/attacking-discriminat... ) isn't particularly compelling. As their example shows, there're two traditional approaches: maximize profit and group-unaware. * Maximizing profits gets the most profit, but treats people differently based on their group. * Group-unaware treats everyone the same regardless of their group, but can generate far less profit. The example pre…

I appreciate the detailed object-level analysis of that case, but my point was more intended at the meta-level — in order to have a conversation about AI bias, we need language and examples to start from. Google is building these fairness metrics into parts of its cloud ML toolchain, and is investing in peer-reviewed research in this area. That’s more than most other companies, and should be applauded. At the object…

> My definition of a sound-byte is a conceptually thin phrase that sounds good, but does not contain much content.

Yeah, same. And to my eye, this whole thing's paper-thin; it tested my suspension of disbelief to not dismiss it as an obvious joke or bad attempt at a PR stunt, and I'm honestly still undecided.

Part of the issue is that its conclusions are absurd. We can invent ethical constraints to arrive at max-profit by pulling things out of thin air; the procedure's analogous to p-hacking. For funsies:

1. Come up with a huge portfolio of ethical-sounding constraints. Include basically any ethical-sounding argument imaginable, even if it seems stupid or bad for business (the numbers can be hacked; we just need the pretty words).

2. Come up with model-transform strategies. For example, instead of calculating the threshold, calculate the log-threshold or the threshold-entropy. (Again, don't think logically -- we're hacking the math, so just come up with random things.)

3. Come up with model-hybridization strategies. For example, construct a meta-model that uses 50% of the threshold + 25% of the inverse-log threshold + 20% of the inverse-entropy threshold + a flat 5% (because, hey, why not?).

4. Whenever we want to maximize profit, just do so with simple profit-maximization. No "ethical" constraints.

5. Repeat Step (4) for millions (or more) different hybridized models, combining all sorts of different model-transforms and ethical-constraints at random. Optimize each for approach to simple-maximization.

6. Stop upon finding an "ethical" solution that effectively equals simple profit-maximization.

7. Generate a description of the "ethical" solution, and write a report on how your company's using that particular ethical strategy as part of its latest, on-going campaign to promote social equity.

8. Just do simple profit-maximization. Or the "ethical" solution; whatever, they're literally the same thing. But if anyone asks, be sure to tell them it was for ethics.

While we're at being ethical, we could even shift some paradigms to envision new modes of engagement with socioeconomically disadvantaged stakeholders to promote a synergistic realignment between our commitment to ethical outreach and disrupting the zeitgeist with our social media blitz to rapidly accumulate surprisingly liquid social capital, indelibly scrawled across the human blockchain of love and mutual respect, validated in the proof-of-work of billions of sapient hearts yearning to emerge from the trials-and-tribulations of our modern era into the welcoming bosom of a new tomorrow.

But who cares about pointlessly defining jargon for logical inconsistencies, though? It's hallow and meaningless; at most, it might fool people before they realize that it's logically inconsistent, but that'd just be lies. It doesn't seem to have an actual use.

Re: Google offers to help others with the tricky ethics of AI

#32

> Google’s new offerings will test whether a lucrative but increasingly distrusted industry can boost its business by offering ethical pointers Isn’t google a big part of why there is so much distrust to begin with? Seems like asking your oxy dealer for detox advice.

I feel sure that the US military has access to every bit of code at Deep Mind. Although they seem to publish most everything anyway. Just not always with all the code.

That's literally the "killer" application that 100% has obsessed many military planners ever since they saw the Atari playing demos.

And whether people recognize it or not, those are the killer robots. They are built into fighters, defense and planning systems.

Regardless of any objections or how many people sign letters against killer robots, the military strategists have no choice but to include the most sophisticated software available.

Now I assume that there is some "failsafe" in most of not all operational systems that prevent actions that are not authorized. But it is also necessary to make that authorization fairly vague in some circumstances. And there is quite possibly a way to turn that failsafe off in the case of a war or scenarios where they feel there is no time to wait for the human (which seems guaranteed assuming the other side has AI also).

Re: Google offers to help others with the tricky ethics of AI

#33
post #19

A lot of knee-jerk Google-hating in this thread, which I think is unfounded in this specific context. Google’s AI safety and AI bias toolchain is by far the most robust that I’ve seen. Seems to me that they are investing much more heavily than other players. Note that this is a different issue than data privacy, which Google rightly takes flak on. AI bias is referring to questions like “if I train a network to (sente…

The issue is that Google has lost most of it credibility as an authority on ethics. Think about what would prompt one to add "don't be evil" to their company charter, then think about what would prompt one to remove it.

Re: Google offers to help others with the tricky ethics of AI

#34
post #26

Earlier quoted context omitted.

Facial recognition will be a far bigger problem when it works perfectly.

Sure but biased AI disproportionately affects people of color which is very bad already.

Why was this downvoted? It's absolutely true and a big topic in AI fairness and ethics work.

Re: Google offers to help others with the tricky ethics of AI

#35
post #34

Earlier quoted context omitted.

Sure but biased AI disproportionately affects people of color which is very bad already.

Why was this downvoted? It's absolutely true and a big topic in AI fairness and ethics work.

Honestly I don’t know if it’s the algorithm or I’m just becoming more radical than the average HN person but I’m seeing more and more weird/senseless downvotes on my comments lately. Makes me want to find another place to hang out. If people here don’t care about marginalized groups then I should be somewhere where they do.
Post reply on HN