Tech giants agree to child safety principles around generative AI

23 April 2024, 15:34

Child online safety report
Child online safety report. Picture: PA

Amazon, Google, Meta, Microsoft and OpenAI have signed up to the safety commitments, which are being led by child online safety organisations.

Some of the world’s biggest tech and AI firms have agreed to follow new online safety principles designed to combat the creation and spread of AI-generated child sexual abuse material.

Amazon, Google, Meta, Microsoft and ChatGPT creator OpenAI are among the companies to have signed up to the principles, called Safety By Design.

The commitments have been drawn up by child online safety group Thorn and fellow nonprofit All Tech is Human and sees the firms pledge to develop, deploy and maintain generative AI models with child safety at the centre in an effort to prevent the misuse of the technology in child exploitation.

The principles see firms commit to develop, build and train AI models that proactively address child safety risks, for example by ensuring training data does not include child sexual abuse material, as well as maintaining safety after their release by staying alert and responding to child safety risks that emerge.

Generative AI tools such as ChatGPT have become the key area of development within the technology sector over the last 18 months, with an array of AI models and content generation tools being developed and launched by the major firms.

The more companies that join these commitments, the better that we can ensure this powerful technology is rooted in safety while the window of opportunity is still open for action

Dr Rebecca Portnoff, vice president of data science at Thorn

The rapid rise has seen social media and other platforms flooded with AI-generated words, images and videos, with many online safety groups warning of the implications of more fake and misleading content being seen and spread online.

Earlier this year, children’s charity the NSPCC warned that young people were already contacting Childline about AI-generated child sexual abuse material.

Speaking about the new agreed principles, Dr Rebecca Portnoff, vice president of data science at Thorn, said: “We’re at a crossroads with generative AI, which holds both promise and risk in our work to defend children from sexual abuse.

“I’ve seen first-hand how machine learning and AI accelerates victim identification and child sexual abuse material detection. But these same technologies are already, today, being misused to harm children.

“That this diverse group of leading AI companies has committed to child safety principles should be a rallying cry for the rest of the tech community to prioritise child safety through Safety by Design.

“This is our opportunity to adopt standards that prevent and mitigate downstream misuse of these technologies to further sexual harm against children. The more companies that join these commitments, the better that we can ensure this powerful technology is rooted in safety while the window of opportunity is still open for action.”

By Press Association

More Technology News

See more More Technology News

Exclusive
Jordan Stephens, Rizzle Kicks star.

Rizzle Kicks star says children 'rely' on online communities for connection as he says 'boredom' to blame for rising crime

A message on an iPhone

Media denied entry to tribunal thought to be about Apple and Government data row

Education Secretary Bridget Phillipson (PA)

Disruptive phones have no place in schools, Education Secretary says

A finger hovering over a phone screen with the Facebook, Instagram and WhatsApp logos

Fact check: Hoax posts about killers and stabbings in local Facebook groups

A drone in the air with countryside behind

Drones used to sow tree seeds in scheme to restore lost South West rainforests

ASCL president Manny Botwe

Technology ‘being weaponised’ against schools and teachers – union leader

A woman using a laptop as she holds a bank card

Phishing campaign impersonating Booking.com targeting UK hospitality

Crypto regulation

NCA officer charged following alleged Bitcoin theft

Sir Keir Starmer walking out the door of 10 Downing Street carrying folders under his left arm

Starmer’s plans to shape up ‘flabby’ Civil Service could trigger union clash

A person holds an iphone showing the app for Google chrome search engine

Apple and Google browser dominance harming consumer choice, says watchdog

A. Lunar Eclipse, Red supermoon, Blood moon / 
on 28th September 2015.

Blood moon 2025: Rare lunar eclipse to be visible in the UK this week - here's how to see it

Several customers took to X to discuss their situation (PA)

Vodafone customers report internet problems in CityFibre outage

Children in school

No nationwide smartphone ban in Welsh schools, report recommends

TikTok is set to launch new parental monitoring tools.

TikTok to launch new parental monitoring tools as app sets limit for teens

TikTok on a phone

Prescriptions for ADHD drugs jump 18% year-on-year, figures show

Elon Musk said there was a cyberattack on X (Kirsty Wigglesworth/PA)

Pointing finger at Ukraine after X outage is ‘dangerous’, cyber expert says