James Titcomb

While artificial intelligence has dominated headlines in the world of technology this year, James has shone a light on the very real dark side of AI.

In 'Why computer-generated child abuse is next crime wave waiting to happen', James investigated the rise of harmful imagery generated by AI. He worked with researchers and experts who detect illegal content to reveal how the freely available AI software Openjourney was already being used to create deep fake illegal images which were being shared on online forums. The article revealed that the National Crime Agency was working with technology giants on shutting down computer-generated illegal images, and a child safety worker raised concerns that the pictures will be almost impossible to track online.

After the investigation, websites hosting Openjourney restricted the tool to block certain keywords, limiting the software’s use.

James also spoke to the Internet Watch Foundation, the UK’s hotline for reporting child abuse imagery. In 'Flood of AI imagery risks overwhelming Britain’s defences against online child abuse', the body warned for the first time that it risked becoming overloaded as criminals increasingly used AI tools to generate abuse material.

The IWF later wrote to Rishi Sunak warning about the surge in images being created by AI and urging the Prime Minister to tighten up legislation to address it. The Government subsequently confirmed that the Online Safety Bill would cover computer-generated images.

James has also investigated the growing trend of relationships with synthetic girlfriends and boyfriends on AI chat apps in ‘A relationship with another human is overrated – inside the rise of AI girlfriends'.

The longform piece used exclusive data on how widespread AI dating is, personal reporting, and interviews with both users and experts to examine the appeal of apps such as Replika. The essay skilfully weaved together the history of humans’ relationships with computers, real-life examples, psychological warnings and an explanation of where technologies such as deepfakes and improvements in machine learning might take relationships.

It came in the same month that a court heard that the intruder who planned to assassinate the Queen in 2021 had been encouraged to do so by his AI girlfriend.

With the Government hosting a safety summit in November focused on addressing the existential risks from AI, James’ reporting has helped bring the abstract concerns about the technology into the real world and illustrated the need for addressing it today.