SK Telecom invokes Minority Report with 'AI crime prevention'

The partnership is fairly broad in scope. To begin with, the telco will work with the university's Security Policy Research Institute to figure out whether AI can be used to curb advanced financial crime – presumably this covers things like fraud and embezzlement, and so-on.

Once they've thwarted that group of ne'er-do-wells, the two will establish a "mutual support and cooperation system," under which they will collaborate on AI technology and R&D verification. They will also work together to identify new areas of joint R&D and technical cooperation.

They also plan to – and this is where things could take a nasty turn for privacy and personal freedom – expand the scope of technological cooperation, "such as incorporating vision AI into CCTV containing crime scenes."

"We hope that by combining the R&D capabilities of SKT and the National Police University's Public Security Policy Research Institute, we will be able to help solve increasingly intelligent security problems," said SKT CTO Yang Seung-hyun. "We will contribute to protecting the safety of citizens by utilising cutting-edge ICT technologies, including AI."

Lee In-sang, director of the National Police University Public Safety Policy Research Institute, added: "We expect this MoU to play a big role in further solidifying the safety of the people."

"We will derive practical results through continuous cooperation and research with SKT and apply those results to public safety sites."

That sounds great, provided there is sufficient oversight to ensure that whoever is in charge of administering this powerful technology does not get too ambitious and exceed their mandate. The road to Hell is paved with good intentions, and AI could prove to be not so much a road but an eight-lane highway – particularly when you start letting government authorities play around with it.

Meanwhile, SKT's fierce local rival LG U+ has announced its arrival to the generative AI (GenAI) party.

It plans to launch no fewer than eight generative AI (GenAI) services within a year, using its in-house large language model (LLM), Exigen.

Technically, Exigen is a small large language model (sLLM), a pared-back, more economical version that is less taxing on compute resources.

Exigen will serve as the foundation for LG's GenAI services, which confusingly enough, will bear the name Ixigen, and will include things like chatbots and retail advisors and so-on. These will be used not only by LG to improve its own services and processes, but also pitched at corporate clients that want to do similar.

LG also unveiled 'Exi Solution', an AI development platform that clients can use to create and test their own GenAI applications.

So far it sounds similar to the countless other GenAI announcements doing the rounds these days.

But the key differentiator that LG is keen to emphasise is the smallness of its LLM. It means that it can be deployed on-premises, lending itself to highly-regulated industries that have to tread carefully with their data – industries like finance, manufacturing, and public administration.

"Ixigen's strengths are 'lightness and speed,' and by applying Ixigen to various in-house services, we are helping to innovate the digital experience of not only our employees but also our customers," said LG U+ chief data officer Gyu-byeol Hwang. "In the future, LG U+ will increase the use cases of Exigen and Exi Solution, advance AI technology and accelerate AX (AI conversion) innovation, while also speeding up business expansion into the B2B area."