Is AI lying? (And other reader questions, answered.)

For the last few years, we’ve been asking Future Perfect newsletter readers what their biggest questions are. And while we usually answer privately, we figured we’d try something new: a reader mailbag!

This week, we’ve answered questions from three readers on classic FP issues: artificial intelligence, animal welfare coverage, and, of course, altruistic kidney donations. We’d like to do more of these, so if your question wasn’t featured — or privately answered — please stay in touch for a chance to be included in the future.

We’re also kicking off the process for our annual Future Perfect list of changemakers. We’re looking for experts, humanitarians, activists, movers, and shakers in global health, broadly speaking.

If there is someone you want to nominate, a topic you want explained, or a question you want us to answer in the future, fill out this form or email us at futureperfect@vox.com. — Izzie Ramirez, deputy editor

By which methods can one ascertain that whatever is produced by AI is exact and truthful?

For any question you’re considering asking an AI model, the first thing you need to do is think about its epistemic nature: Is the answer knowable in an objective way? Or is it subjective?

The best use case is a situation where it’s hard for you to come up with the answer, but once you get an answer from the AI, you can easily check to see if it’s correct. I find chatbots particularly helpful for semantic search — that is, cases where I say, “There’s some psychology theory or idea in philosophy that basically says XYZ, but I can’t remember what it’s called or who said it, help!” The chatbot will give its best guess, and then I can just fact-check that.

Same with other empirical facts that are verifiable through observation or data — anything from “What’s the boiling temperature for water?” to “Is it true that humans share 98.8 percent of their DNA with chimpanzees?” While you can easily verify the first by yourself through observation, you’ll need to rely on experts’ data for the second. In that case, you need to feel confident that what’s produced by your fellow humans is exact and truthful. We’ve developed tools that increase our confidence, like the scientific method, so if you’re consulting scientific experts, you can at least have some degree of confidence that they’re reporting observable and repeatable facts.

Then there are domains that are inherently subjective. If you’ve got the type of question for which there is no One True Answer, you’ll want to be more hesitant about using AI. I think ethical dilemmas fall into this category; no matter how much OpenAI tries to create a “universal verifier,” AI will always be limited in its ability to advise you on how to handle an ethical dilemma, because there’s no One True Ethics. So, you might see what thoughts an AI model provokes in you, but don’t trust it as giving you the final answer, especially if what it’s saying seems off to you. In other words, you can use it as a thought partner, but don’t treat it like an oracle.

— Sigal Samuel, senior reporter

Ok, after more than five years as a vegan and 73 years on the planet, I want to know why the great majority of journalists consistently abandon everything they learned about objectivity when it comes to a multitude of issues with the monster industry known as “animal agriculture?” And I want to know how to combat that bias effectively.

It is a huge blind spot for most of them. My best guess is the conditioning is so strong. It starts as a toddler, is reinforced by the parental relationship, expands to extended family, friends, reinforced again by all types of advertising media, entertainment, etc. Then they go to journalism school and are taught by instructors who also have this blind spot.

So later a reporter will go to a “chicken farm” and empathize with them when they tell their story about losing thousands of birds to avian flu — their sense of loss is not about the birds; it’s about the money. The reporter presents the story without questioning the basics. Things like “where are all the male birds?” [and] “how is it possible for anyone to think that 35,000 birds could be forced to live together in a building without reasonable access to the outdoors?” and “why does it smell so bad?” and “why do you have permission to confine animals without their permission?”

I think the reason is pretty simple: Journalists are people with their own biases, just like everyone else. That’s evident in how little coverage factory farming receives in the first place — it involves the abuse of billions of animals and hundreds of thousands of workers, and is a leading cause of many of our environmental problems, yet only a handful of US journalists write about it full-time (including yours truly). Most news outlets and editors don’t take factory farming seriously, which is why I’m proud to work at Vox, where we do.

That’s the most fundamental problem. But secondarily, while there is plenty of fantastic coverage of factory farming, more often than not, I find I’m disappointed with a lot of it, too. I see a few recurring issues:

I’ve written one story about how the media could cover these issues better, and I hope to keep covering that in the future.

— Kenny Torrella, senior reporter

Stories like Dylan Matthews’s years ago led me to investigate donating a kidney to a stranger. I asked my doctor about it, and surprisingly, instead of encouraging me to save a life, he tried to talk me out of it.

He told me that it is illegal to donate a kidney to a stranger! I live in Hong Kong, and maybe the reason for prohibiting even the donation of a kidney to a stranger is the fear that people would secretly accept payment from the kidney recipient. But I don’t know why. Anyway, I thought about donating while on a vacation in the US, but it would require too much time, so I gave up.

Unfortunately, my second kidney will probably die with me in old age, and someone with kidney failure will needlessly die. Anyway, maybe another story idea would be about paying kidney providers in countries other than the US?

Most people aren’t as generous as you!

In the US, only a sliver of living donations go to strangers. Meanwhile, over 100,000 people sit on kidney waitlists. And, as you indicate, the need for kidneys is a global problem, too.

Many places only allow donations to relatives or known recipients (or require tough ethics reviews for unrelated donors), while a minority — like the US, UK, Canada, and Australia — offer a formal pathway for anonymous “good Samaritan” donors. In Hong Kong, where you’re based, you can donate to a family member easily, but unrelated donations need official approval, and there’s no standard program for that. (That’s probably why you were discouraged.)

This patchwork exists for a reason.

In the 1990s and 2000s, there was a serious trafficking and transplant tourism problem. In 2007, the WHO estimated that about 5–10 percent of kidney transplants involved trafficking, and countries like the Philippines and Pakistan became hubs for foreign patients buying organs from desperate locals.
Transplant experts met in Istanbul in 2008 and wrote what became the worldwide rulebook. The Istanbul Declaration pushed countries to crack down on coercive sales of organs. Every country had its own laws, but began incorporating the declaration’s recommendations. As a result, transplant tourism dropped sharply in Israel and the Philippines once new rules kicked in, and tighter oversight became the norm across Europe.

But, in its efforts to shut down trafficking, the declaration argued that compensating donors at all “leads inexorably to inequity and injustice.” There was little empirical data to back that claim, but because it came from a major international statement it hardened into gospel: organ donation must be “financially neutral.”

But neutrality isn’t actually neutral in practice. Living donors lose wages, take time off work, take medical risk, and sometimes even face higher insurance premiums after donating. We don’t call that exploitation — but it is a penalty for doing the right thing.

And it’s inconsistent with how we treat other socially valuable, risky, or unpleasant work. We pay people to do jury duty. We pay clinical trial participants. In many places, we even pay plasma donors.

There is one striking exception: Iran.

It’s the only country with a regulated system that pays kidney donors. Iran established this system in 1988, and today performs about 2,500-2,700 kidney transplants annually, and it claims to have essentially eliminated its waiting list. It’s a proof-of-concept that incentives can be structured.

The US debate is inching in that direction. Congress’s End Kidney Deaths Act would offer a federal tax credit to people who donate a kidney to a stranger. Donors would receive a $10,000 tax credit annually for five years, so not quite direct payment, but certainly a help. The act, which has not been voted on yet, acknowledges that donation involves real costs: time off work, medical risks, recovery time.
The path forward globally isn’t throwing out Istanbul’s anti-trafficking work, but to build on it with smart incentives and guardrails so people can donate altruistically if they want to. That means actually testing new approaches, but doing it carefully. Give donors independent advocates, make sure there’s time to think it over, and guarantee lifelong follow-up care.

In the meantime, you might not be able to easily donate your kidney to a stranger right now in Hong Kong, but the needle is moving in the right direction.

— Pratik Pawar, Future Perfect fellow

Want more Future Perfect? Sign up for our newsletter here.