I would say that it's far more likely that the system simply never worked properly. The amount of non-functional so-called "AI" features you find in China is astounding. A friend's apartment building got the latest facial-recognition door lock - which will open the door if you hold up his Facebook photo on your phone. Just the other day, my wife found out by accident that her friend's Huawei phone will unlock for her using its version of Face ID. They have similar facial structure but really look nothing alike. China is amazingly good at many things, but _genuine_ technological innovation is not something I've seen in my time here.<p>No doubt this corruption-busting "AI" was developed by companies and people with deep connections to the relevant Party members, but little of the necessary experience or resources to actually carry out the project successfully. When it became clear that it didn't work and would never work, this story was a way to allow the responsible people to save face.
“AI may quickly point out a corrupt official, but it is not very good at explaining the process it has gone through to reach such a conclusion,” the researcher said. “Although it gets it right in most cases, you need a human to work closely with it.”<p>And that's the crux of the issue with AI being used in any law enforcement situation.<p>If we allow decisions and conclusions being drawn by the AI without a clear explanation of how it got there, we're just creating a monster that will advise -maybe replace- the judgment of law enforcement professionals who won't have the means to question these decisions.<p>Catching corrupt officials is a laudable goal anywhere and I'd like to see it applied to highlight potential irregularities that may require a second look but the danger here is that an unprovable AI be used to make claims or start being used as sufficient evidence to ruin people's lives.
This reminds me of a detail from the plot of the original Deus Ex. Part way through the game, there is an optional dialogue with a benevolent AI. If you question it enough, it reveals that it is the rejected prototype for an anti-corruption/terrorism AI which was turned off for being too effective: eventually classifying its own masters as a criminal organisation.
"it is not very good at explaining the process it has gone through to reach such a conclusion".<p>The above statement is just too convenient and practically superstitious.<p>AI is data hungry. That implies that there are so many past incidences of corruption you can generate large training data sets. So perhaps even a random guess would be <i>too efficient</i> since it is right more often than not.
"Everyone has committed a crime, it's about who we decide to prosecute" --KGB <a href="https://www.wsj.com/articles/SB10001424052748704471504574438900830760842" rel="nofollow">https://www.wsj.com/articles/SB10001424052748704471504574438...</a>
The 'problem' with widespread corruption is not identifying it, it's always been an issue of the political will to do something about it.<p>When it's rampant, it's easy to find.<p>So this has nothing to do with AI or even technology. Granted, tech might make it easier to find some bad guys, even then, the fact that stuff is 'online' and in a 'searchable DB' makes this possible, the AI is not necessary.
This sounds like a really typical implementation of an AML system. My guess is they actually shut it off because it generated too many false positives. In authoritarian countries, corruption charges are generally trumped up when an official becomes inconvenient for those in power. This means that there are a lot of things officials do that are tolerated and will be used against them once they need to be removed. Therefore the definition of corruption is loose as it's kind of an accepted operating model. If you implement a system like this, suddenly all that acceptable behavior generates lots of positives because your feature selection and classification is based on a threshold no one meets.
Here in Kazakhstan a couple of wise guys suggested to use a blockchain to fight crimes within the police itself, and these plans quickly ditched after they realised the blockchain's pure brutality.
Europe has a similar system in place.<p><a href="https://ec.europa.eu/social/main.jsp?catId=325&intPageId=3587&langId=en" rel="nofollow">https://ec.europa.eu/social/main.jsp?catId=325&intPageId=358...</a>
"AI" might be as simple as running a calculation on his expenses, real estate holdings, $$ deposited, wired out, salary etc. Maybe run for his close family, driver's family too.
High level issue
================<p>Given how corruption is being used and how low pay the top guy is, i think the AI need HI very much.<p>Put it the other way everyone must be corrupted somehow otherwise how can a us$100k top communist sent her daughter to harvard to study psychology. And the other guy who died of corruption charge run a fast railway wort trillion dollars for petty salary.<p>They have to reform the compensation package ... communist have a hard time here. Hence has to lie ... and if a system can spot liar ...<p>Muslum jailing issue
====================<p>I heard this black ... but both should not be done and given there is not even fake news independence and judicial system, stange to compare
> Beijing has been developing a nationwide facial recognition system using surveillance cameras capable of identifying any person, anywhere, around the clock within seconds.<p>So before all the privacy activists are up in arms, this is pretty incredible and it looks like they're getting pretty close to eliminating all violent crime. I think that's an incredible achievement if they can pull it off. In an even broader historical context, individualism and capitalism have had their run for 100+ years, maybe this is the rise of a new ideological movement.