Online platforms and social networks including Google, Facebook, and YouTube should be held to tougher transparency and accountability standards when it comes to targeting content at users, UK advisors say.
The Centre for Data Ethics and Innovation (CDEI) is an advisory body established under the Department for Digital, Culture, Media & Sport to provide independent advice and recommendations on emerging data-driven issues including AI, deepfakes, surveillance, and Internet of Things (IoT) technologies.
On Tuesday, CDEI published a report and set of recommendations (.PDF) on content targeting -- the use of user data, patterns, and activity in targeted advertising, as well as to recommend sponsored content such as images, posts, and videos -- on online platforms.
AI has entered the fray as the means to create machine learning (ML)-based algorithms that collect and analyze our behaviors, interactions, likes and dislikes, searches, and connections can be used to influence what we see through our online experiences.
After consulting with industry experts, conducting interviews, communicating with eight regulators, a series of public engagements, and launching open calls for evidence, the advisory body's final report suggested that personalization can increase the usability of the Internet, improve navigation, and is a driver of "economic value" for business models employed by some of the world's largest companies.
Personalized services, too, can help users improve their own lives through fitness and activity monitoring, financial assistance and control, and there is "significant potential" for innovation in the sector, according to the CDEI.
It is expected that AI-driven targeting will continue to develop and expand in the future, but therein lies the risk -- as the CDEI says that online targeting systems are all too often are implemented without transparency or accountability, falling short of human-central AI principles created by the OECD, of which the UK has subscribed to.
"Online targeting has helped to put a handful of global online platform businesses in positions of enormous power to predict and influence behavior," the report says. "However, current mechanisms to hold them to account are inadequate."
The advisory body also raised the topic of harm caused by data-driven content recommendations; such as the exploitation of human vulnerabilities, bias, and discrimination.
The CDEI has made a number of recommendations to hold online platforms and services more accountable based on its findings.
The first is systemic regulation when recommended content like posts, videos, and adverts are involved; but instead of specific restrictions or laws, CDEI has suggested the creation of a regulatory framework which "promotes responsibility and transparency and [to] safeguard human rights by design."
CDEI has also proposed the reform of how users can control how their online experiences are personalized, while also maintaining rights to free expression, and the creation of online, public archives of "high-risk" adverts such as political ads and age-restricted products for information-gathering purposes.
New powers, too, have been suggested to allow independent researchers secure access to their data to build an evidence base on issues of public concern such as data protection compliance, AI bias, and fake news trends.
A proposed new regulator for online harms should be established, CDEI says, and should work with the UK's Information Commissioner's Office (ICO) and the Competition and Markets Authority (CMA) to work with online platforms to maintain transparency and accountability standards.
"Most people do not want targeting stopped," the CDEI said. "But they do want to know that it is being done safely and ethically. And they want more control."
The UK government will have to respond to the CDEI's recommendations within six months of the report.
Previous and related coverage
- Cybersecurity 101: Protect your privacy from hackers, spies, and the government
- Cybersecurity: A guide for parents to keep kids safe online
- US job seekers scrub their social media accounts to get success
Have a tip? Get in touch securely via WhatsApp | Signal at +447713 025 499, or over at Keybase: charlie0