Is AI risk associated with 2024 U.S. election candidates?

Many EAs are very worried about AI risk.  Could different voting habits in the 2024 POTUS election be associated with this? Here is 

Which cities give me a lower chance of dying in a catastrophe?

I remember when the Ukrainian war reached peak nuclear threat.  And I remember some EAs were interested in fleeing their city, for safety.  That inspired me to 

.  It is meant to supplement your decision of "which city should I live in?".

Is stimulant medicating ADHD worth doing long-term?

We need to know if stimulant medicating continues to "work" for ADHD.  We don't have the types of studies that would confirm this.  "If I am medicated, will it actually help me long-term"?  This market is a narrow attempt to illuminate that.

Which crypto exchange could I use to reduce my chances of losing the money, to exchange fraud?

I can only assume this should be of great relevance to a lot of people, after FTX.

Manifold is fucking up again and won't let me search for the embedable market, so I'll post it later.  Series can be 

I've tried sending a crypto exchange whistleblower market to people, and these people didn't want to hear to hear about it.  After FTX.

I think the most misleading objection is that Manifold markets need many traders to be accurate.  

An actually-good objection can be about the criteria.  It's hard to think of unambiguous+relevant criteria, for complex decisions.  But if someone gives me better criteria, I will be happy to add new markets.

Decision markets that should influence, but probably won't

An attempt to give a clear comparison, of which crypto exchanges are relatively safe from fraud.  I'd like to see 5+ traders, but if you think Manifold markets need hundreds of traders to be useful, 

.  If you think your personal judgment will be more-reliable, 

.  If you have an exchange you'd like me to include, please suggest.

Crypto Exchange Risk

A whistleblower group about the crime risks of billionaires.  Part of a 

Billionaire Crime Markets

For criminality markets about the Forbes "30 Under 30" list (more like 600 under 30).  Part of a 

ForbesUnder30 Criminality

For markets that ask whether a study finding will replicate.

Best practices: the "replication" should expect more statistical power (or participants, etc) than the original.  Or there should be some other enhancement to the method.  Otherwise it's too easy to get spurious "replications".

Replication Markets

A criminal charge goup about people who are CEOs of major companies. Part of a 

CEO Criminality Markets

A criminal charge goup about people who are running for the 2024 U.S. presidential election.  Part of a 

About post for the group

POTUS 2024 Candidate Criminality

For the charge markets that are about people in crypto, or associated with cryptocurrencies.  You can see a pre-emptive 

CryptoChargeMarkets

For the subset of charge markets about people who are related to Effective Altruism.  Includes some people who don't identify as EA and are only very indirectly related.  Results so far are notably much lower than my crypto markets.

EAChargeMarkets

Why is there a crime market about [person]?

Most smart people got a good impression of SBF.  That’s very instructive.  If you only make crime markets about the visibly-risky, you're massively under-using them.  My question sets must be impartial and broad.  For example in the U.S. election set, I must include both candidates who are like Trump, and also ones who are very different.

But most of the people in my markets will never be charged of anything.

When should these prices be taken seriously?

I doubt there's some accuracy cutoff (I guess it scales lognormally).  But as a rule-of-thumb, I recommend ignoring the price signal until there are 5+ traders.  Since I know the creation price (50%) is generally too high, I place a small "NO" position on each new market.  Partly to help push it closer to the base rate, and partly as a gesture of goodwill.

What is a "serious crime", and/or "criminal charge"?

Some things that count: wire fraud, perjury, murder, tax evasion, assault, driving under the influence, and robbery.  This will vary by country, but in general, I'm going to ignore "small possession of marijuana".  The intention is to focus on crimes with an actual victim, and ignore personal indiscretions.  Jaywalking won't count, unless the person was endangering others (such as by obstructing traffic).  Littering generally won't count, unless it was an unusual amount, and they received a serious punishment.

They must have been charged by the police, or a grand jury.  Note that in the U.S. this 

include SEC "charges", which are actually lawsuits and not chargesh.  Though SEC charges can often lead to criminal charges, depending on the case.

Note: the earlier question wording of "criminal charge" was too U.S.-centric.  So newer questions mention "serious crimes", and I've tried to elaborate with some examples.

 of my markets are about people who were born in authoritarian countries like China.  This may be hard to resolve unfortuntately, if they are imprisoned without a publicly-stated reason.  So 

, if they are detained or interrogated but then released without a "criminal record", then that won't count.  If they are imprisoned for 3+ months but still no due process, we may be forced to resolve "YES".  It's unjust, but such a duration is long enough to jeopardize the persons' business interests or enterprise.  So while unfair and very staining of that country's business climate, we would learn towards resolving YES, because rewarding the market for predicting long-term imprisonment is useful.

I see my market, and I think the price is wrong!

Then make a Manifold account and trade it to whatever price you think it should be.  You're allowed to "insider trade" -- on average that helps the price get more accurate.

For easier interpretation, they should all have the same timeframe.  If there is any institutional interest in these, then I'll someday make a 2040 series.

Are you going to make these markets about everyone?

The world still can't reliably avoid giving billions to someone like Sam Bankman-Fried.  That seems alarming to me.

If there were markets on everyone, then it would be normal.  Like credit scores.  But I don't have the capacity to add more than a couple per day.  I'm prioritizing people who seem important, and getting enough mana to pay for itself.

Why bother with decadal prediction markets on crime, when GPT-5 might just go build a dyson sphere?

I see this attitude on Twitter.  It's such a bad forecast; fraud will probably matter for many years to come.  It's remarkable for me to hear this, so shortly after Sam Bankman-Fried.  Anyway, if you have an AI that can estimate crime risk, and is SO GOOD that it credibly outperforms prediction markets, then please let me know.  I suspect I'll be waiting many years though.

I can think of important people you should have markets on.