Transcript
Julie Chang (0:00)
The Apple Watch Series 10 is here. It has the biggest display ever. It's also the thinnest Apple Watch ever, making it even more comfortable on your wrist whether you're running, swimming or sleeping. And it's the fastest charging Apple Watch, getting you eight hours of charge in just 15 minutes. The Apple Watch Series 10, available for the first time in glossy jet black aluminum compared to previous generations. IPhone XS are later required. Charge time and actual results will vary.
Julie Chang (0:28)
Foreign welcome to Tech News briefing. It's Thursday, January 23rd. I'm Julie Chang for the Wall Street Journal. Could artificial intelligence learn to speak financial jargon? Compliance companies say they've cracked the code and are using the tech to fight insider trading and other financial crimes. Plus, screens are becoming a bigger part of the school day in classrooms across the country, accelerated by the COVID pandemic. But has more tech in classes benefited learning? We'll dive into that. But first, catching insider trading tends to be challenging because of all the financial jargon used by Wall street traders. Now companies that make compliance software are pitching generative AI to firms, saying the tools could help combat financial crime. Risk and compliance Journal reporter Richard Vanderford has been looking into this, and he joins me now. Richard, what problem are AI compliance software companies looking to address here?
Richard Vanderford (1:34)
Banks and other kinds of financial institutions create a large amount of communications data, obviously, because they have thousands of employees, and the compliance staff have to sift through that to see if there's any evidence of market manipulation or insider trading or any other kind of financial crime. And in the past, this had been done by frontline staff using pretty rudimentary computer tools looking for maybe a word like insider trading, if someone would say that and then someone manually reviews it. And the promise of AI in this context is that it could simplify and maybe automate some of this work and also potentially make it so that the bank or other institution doesn't have to hire as many compliance staff going forward.
Julie Chang (2:21)
What are some of the jargon or code words that traders use?
Richard Vanderford (2:24)
There's one example I was given. Someone would say something like, let's throw some chum in the water, and that's a reference to spoofing. But a traditional system wouldn't catch it because there's no word in there that indicates like a crime could be occurring. Chum is not something you would want to flag because you'd have way too many false positives. But an AI system might be able to catch this and say this looks like spoofing, or it could be spoofing the Systems can also potentially catch emojis being used or other ways to use code language. As the technology to screen these communications has gotten better, duplicitous traders have gotten more sophisticated about saying things in code, maybe just responding with an emoji. And it would be very hard to catch with a traditional compliance system. But an AI powered system could potentially say, oh, that I think they're actually talking about something suspicious or that indicates a potential financial crime.
