UK frontline images against child abuse warn of AI threat

AI For Business


The use of AI image generation tools has exploded in recent months, enabling users to create photorealistic images in seconds with just a few written instructions.

Most famous programs introduce restrictions on the generation of illegal or pornographic material, but users have shared guides on how to get around these restrictions, as well as “open source” alternatives that don’t have such restrictions. I’m looking.

Activefence, a company that monitors online forums, identified 68 sets of AI-generated child abuse imagery on one website in the first four months of this year, compared with 25 sets in the last four months of 2022. Announced.

There are also concerns that the dataset of billions of photos used to “train” the software contains illegal content. Possessing and distributing fake child abuse imagery is illegal in the UK.

One content moderation executive said AI-generated images created a big problem for law enforcement because they were not recognized by the software used to determine if illegal images had been previously reported.

This means that a large amount of fake images can be treated as brand new, which can lead to false concerns that children are at risk.

A spokesperson for the National Crime Agency said, “We are constantly considering the impact of new technologies such as synthetic media (including those generated using AI) on the threat of child sexual abuse. We are working closely with law enforcement and law enforcement partners.” We want to enlist the broader government and private sector to work with us to ensure our capabilities can continue to detect and investigate synthetic media as technology evolves. ”

Baroness Kidron, a child safety activist and colleague in other organizations, has advocated for AI abuse to be included in the government’s online safety bill, with huge fines for companies that do not engage in harmful content. will be

She said she introduced the fix after being told by police that AI-generated abuse images had exploded this year.

AI firms said this week that policymakers should focus on “extinction” threats on the scale of pandemics and nuclear war, but critics say that’s a distraction from today’s problems. claimed to be.

Rishi Sunak is scheduled to discuss AI with Joe Biden in Washington next week.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *