[geofft]

> My sources at Google just couldn't see the panelists constructively working together on a panel at this point.

Well, that's the point, isn't it? If the AI ethics council is meeting four times a year for a couple hours at most, and and has people who can't even agree on "Which bathroom should this person use," how are they going to produce productive advice for actual hard questions that haven't been well-explored?

There is a place for debate between people who don't agree on worldview. This council was never going to be it.

[qmanjamz]

It seems to me that if they limited their recommendations to only things the entire board could agree on, they could still be productive.

For example, they may not agree on how bathrooms should be organized, but they probably all agree that bathrooms should exist.

The fact that the board had wildly varying opinions is a feature, not a bug. The outputs of this board should be widely agreed upon. If they make declarations that half the country disagrees with, they're going to lose credibility fast. Having an ideologically diverse board helps to secure their credibility.

[geofft]

But the board is there to advise Google on hard questions, not easy ones. Is it unethical, for instance, to use AI to detect and block potential child pornography and flag it for potential referral to law enforcement? Probably not, but anyone can tell you that. Is it unethical to use AI to identify "gender ideology propaganda" and refer that to law enforcement in places where it's illegal? Depends on whether you think such laws are ethical. If the council can't answer that, what use is it?

(If the point of the council is to tell you things that anyone with a functioning conscience could tell you, then the fact that it's needed is a serious indictment of Google management. I'd like to think the council is here for the hard questions.)

[qmanjamz]

If you stack your council with people of entirely one viewpoint, there's little point in having a council at all. Create a board entirely of Muslims and it's going to say we should fast during Ramadan. If Google tries to cite that board as an ethical authority to justify a policy of mandatory fasting, they're not going to convince many people.

With a diverse board, the answers to some questions are going to be inconclusive. But that doesn't mean the board shouldn't exist at all. And it certainly doesn't mean we should eliminate all diversity in the board.

[educationdata]

How many conservatives do they have on the board? I feel viewpoint diversity should at least be close to nationwide percentage.

[geofft]

Which nation? Google operates in many countries - should it be proportional to business activity in each country?

[educationdata]

Is "Which bathroom should this person use" an easy question? If it is really easy, why it is so controversial? You may think it is an easy question, in reality it is an actual hard question.

[geofft]

It's controversial because it's based in matters of fundamental worldview, not matters of unexplored complexity. So, given your fundamental worldview, the conclusion is pretty easy. There's nothing that a council of great thinkers is going to be able to illuminate for us on the question. We know what the arguments are. It's "easy" in the sense that the textbook trolley problem is "easy": there's nothing more to be learned about the question itself. We understand the problem and the arguments on both sides, and a council of great thinkers can pull up Wikipedia for you if you'd like.

Hard questions, as I mean them, are ones where there's a lot of complexity to even reach a potential answer. Google's trolley problems are those where the decision isn't just "flip the switch," it's "should we gather this data" or "should we conduct this experiment" or "should we participate in the market in this way" where there are unforeseen side effects to doing so, and you want a bunch of smart people in a room to try to figure those out and shake out all the complexity in the problem that you haven't even managed to state clearly.

But if you can't decide whether it's better to kill one person by action or five by inaction in the first place, figuring out whether taking/not taking the action has unexpected ethical implications isn't going to get you closer to any answers.

[matt4077]

„Some say you shouldn’t exist. Others say you have a right to live. Hard question“

[tomp]

Oh come on you're making a huge strawman.

Personally, I don't see a point in sex/gender-segregated bathroom at all. The only segreggation that makes sense to me is sitting/standing (which mostly conincides with private/non-private).

But if we want sex-segregated bathrooms (and AFAIK many governments force that), then there's little point if people can just randomly change their mind about which category they belong to. What's the point of categories then?

[TheDong]

> But if we want sex-segregated bathrooms then there's little point if people can just randomly change their mind about which category they belong to. What's the point of categories then?

Categories do not need to be immutable. There are meetups for bikers and redditors and many other categories which anyone may self-identify as belonging to. These people prefer to think of themselves as belonging within those categories for any of a number of reasons, and accepting their choice is generally a decent thing to do, barring a compelling reason to the contrary.

The point of the male/female categories is that a large number of people identify as one of those two genders, and prefer strongly to be recognized as within one of those categories. There are exceptions and it's more complicated than that, but in the common case, that's why the categories have value: people ascribe value to them. If recognizing someone's desired identity makes them feel more comfortable, why shouldn't we? In the case of genders, it's especially important since it's such an important category.

Finally, people switching genders on a whim is a strawman itself. Transitioning your gender officially is a multi-year process, and you face significant persecution and stigma along the way. I don't see why anyone would undergo that unless it helped resolve some acute gender dysphoria.

[tomp]

> Categories do not need to be immutable.

I agree, but you have to ask yourself, what purpose do the categories serve. It seems to me that th epurpose of sex-segregated bathrooms is two-fold: (1) to make people feel more comfortable, because many have weird hangups regarding intimacy and sex, and (2) to keep women safe. Even disregarding (2), it seems to be a choice between inconveniencing a tiny proportion of the population a bit more or inconveniencing most of the population a bit less. In any case, regardless of what bathroom trans people should go in, it doesn't even remotely start to solve other issues (e.g. what bathroom does a dad take his 4-year old daughter in) which is why I think the whole idea is a huge politically-motivated strawman.

[fixermark]

The error you're making in trading off a bit more discomfort for an already-marginalized group against any discomfort for the mainstream is exactly the reason having someone in an AI ethics board who doesn't see being a trans person as valid was a problem.

It's also the reason we need more ethics boards, because it's a common error technical training leads a person to, where we're taught how to optimize for common cases and defend against unusual cases---when the things we use aren't people and don't care if they're considered "common" or "unusual."

[TheDong]

We could talk in circles about the details here. The fact of the matter is that those who harass others in bathrooms are already criminals (regardless of gender), and trans people are not overrepresented in those statistics at all, so it's a strawman to imply otherwise.

The fundamental difference that will not allow us to converse well here is that you are intentionally not accepting that a trans person is the gender they have transitioned to.

It's clear that's the case from, for example, you saying "to keep women safe" and very obviously not including trans women in that group. Trans women are women and also deserve to use the bathroom where they will feel safer.

[93s6oz]

That is polarising. An extreme take. I haven't seen anybody make the argument that trans people should be killed. Just that men should not be able to use women bathrooms. In this case, that's the argument. Turning it into something like that is not going to help anybody.

[TimothyBJacobs]

The OP didn't say the counter argument was that trans people should be killed, but that they shouldn't exist. You are taking the take to the extreme.

Routinely, trans people are treated as an entity that should not exist. Or a disease or condition that should be rooted out. Not a valid identity.

[pjc50]

> I haven't seen anybody make the argument that trans people should be killed

Brunei?

[sadris]

Think their law applies to homosexuals, not transexuals.

[mike00632]

If I were trans I don't think I would take my chances of them making a meaningful distinction.

[matt4077]

It’s polarizing only if you accept the you-should-die part one of the poles, which is sort of the problem I was getting at. I would characterize it more as reductio ad absurdum.

[UncleMeat]

"Should we own slaves?" is an easy question. Yet it was so controversial there was a civil war fought over it. "Is global warming caused by humans?" is an easy question (with the data we have) but it is massively controversial.

[kevingadd]

It's a very easy question, almost all countries have resolved answers for it. They don't all agree, but it was very easy to answer. Many of them have clear, objective standards on how to apply the rules (that people choose to ignore or argue over), though some certainly are determined to make it as fuzzy as possible for their own agendas. You can easily find local regulations addressing this topic pretty easily for most parts of the US and european countries, also some other countries like Japan.

There's also the standard fallback answer for "which bathroom": if you make them all single occupancy, it doesn't matter which one. Sometimes that's the solution used in situations like schools.

So maybe pick a better example of a hard question that seems easy. (Though I get the relevance here, given the decried panelist's history and views)

[gubbrora]

Why should they agree? They could have produced advice of the form "from perspective A we should do x but from perspective B we should do y".

[geofft]

The point of debate is to test ideas by argument and logic and strengthen the strong ones and reject the weak ones. If you just say "Here are our thoughts," then the "what about robust debate" rationale doesn't hold up, and we can get to the direct question of whether James' worldview is moral or immoral, because the people taking the "advice" are going to have to pick either perspective A or perspective B.

And if you're going to pick one in the end, why not get everyone from that worldview? There are lots of Roman Catholic theologians, for instance, that they could have included. I don't personally agree with them either, but they absolutely have a rich history of thinking about ethics, and they mostly also don't believe in trans people. Why leave James to fend for herself?

[gubbrora]

Well in the hypothetical, whoever takes their advice would need to make a synthesis of it somehow. E.g. most perspectives agree so let's do it this way. Or this question doesn't matter that much from most perspectives and according to this one it matters a lot so let's go with that one. Or even just ignoring them from time to time.

[mc32]

Presumably there would be many more issues beside bathroom policies. Moreover one person isn’t s majority among eight.

[yalue]

You're right about all of that, but shouldn't a meaningful board at least make some show of trying to include diverse opinions? The key word of course is "meaningful", and I think you're right on the money that Google wasn't interested in having a panel to discuss meaningful issues in the first place; it seems more likely that they wanted to be able to point to a source of "external" "unbiased" opinions to justify what they were planning to do already.

(Obviously the people who complained about a prominent right-wing figure would just argue something hypocritical like "only the right kind of diversity of opinion", but that's beside the point; the panel was a dumb idea to begin with, coming from a group with so much of a stake in trying to advance AI for the sake of advertising)

[durability]

Diversity is great! Putting someone who advocates for not using AI on an AI ethics board is diversity. Putting someone who advocates for more/less/self regulation is diversity.

Putting someone hostile towards the LGBTQ community and LGBTQ rights? That’s not diversity. That has no place in 2019.

[yalue]

Sure, but if they're not debating that topic, why should it be an issue? Not that I think such a person belongs on the panel, but the reason they're not invited should be, as you said, that they have few or no qualifications to discuss either ethics or AI.

The fact that they don't agree with the others on LGBT rights should only be an issue if they insist on trying to espouse their unrelated viewpoints using the panel, in which case by all means kick them off.

[durability]

Because you can’t look at these beliefs in a vacuum. Someone who’s against LGBTQ rights isn’t against it for the hell of it, they have some underlying reason such as religious ideology. I’m not saying there’s anything wrong with religion, but when you believe that a man and a man together is unnatural and a sin against nature, then I have to wonder what other backwards regressive medieval beliefs they hold as well.

[yalue]

(Note that I'm not trying to straw-man you here, because I know you're only talking about the specific instance of this panel, which I've already agree was a farce to begin with. However,...) I still think it's harmful to society as a whole to try to use someone's views on any given topic as an excuse to refuse to work with them on something unrelated.

Say, for instance, that you're asked by a supervisor to collaborate with a Muslim living in a middle eastern country on writing a piece of software. For the sake of argument, say that this particular person believed that the laws recently enacted in Brunei were in line with the commands of Islam, and therefore just. (For the sake of argument only--obviously not everybody in those categories is like this at all--but clearly at least some must be or we wouldn't have situations like this recent one.) Would it be right to refuse to work with this person on writing say, some python code, when the topic of "sharia law" is never going to be a factor? Furthermore, by your argument, would you be willing to trust this person's judgement on software architecture when they have exhibited a, in your view, extreme lack of good judgement in another field?

Second, you can even take this argument to a higher level. You admit that the opposing party has a reason to believe what they believe, and I'd imagine that you'd even grant that they probably believe that their reasons for their particular position are good reasons. Leaving aside whether the positions themselves are good or not, how do you evaluate whether someone's reasons for holding a particular position are good? After all, a religious person may believe that one day an omniscient being will call them to answer for why they failed to keep holy commands in their life. To the holder of such a view, that's a pretty good reason to do some twisted stuff. If anything, (if their religion is right) such a reason for them holding their particular position (fear of eternal condemnation) may even be better to them than your reasons for your position are to you!