The ODI has joined forces with Full Fact and international fact-checkers to use artificial intelligence to dramatically improve and expand the global fight against misinformation, having won the Google AI Impact Challenge
Information that is false – or ‘misinformation’ – and false information spread deliberately to deceive – or ‘disinformation’ – affect millions of people’s lives; their health, safety and ability to participate in society. In recent years, we have seen people die in acts of violence fuelled by rumours spread via social media, and a new outbreak of measles, among many other things.
Tackling misinformation is complex and requires people like journalists and fact-checkers to be able to respond at the speed and scale of the internet. In the past few years, new technological solutions have been proposed by academics and fact-checking organisations such as Full Fact, to help tackle these challenges.
These are promising, but risky. Building technology to automate or speed up responses to misinformation requires a deep understanding of public debate. It also needs to be developed in ways that care to protect free speech and pay close attention to the responsible limits of artificial intelligence (AI) in this field, and how these can vary across countries, languages, and social and political contexts.
The Open Data Institute is joining forces with reputable fact-checkers and pioneers in automated fact-checking: Full Fact, Chequeado, Africa Check, to advance these efforts. Together, we will work with media outlets, civil society, platforms and public policy makers worldwide to help them understand how AI can help people decide what information to trust, and bring the benefits of automated fact-checking tools to everyone.
The project – which will use AI to dramatically improve and expand the global fight against misinformation – was announced today as one of just 20 international winners of the Google AI Impact Challenge, chosen from more than 2,600 nonprofits, social enterprises and research institutions around the world.
At the ODI, our vision is for a world where data works for everyone. To successfully tackle misinformation, it’s crucial to have access to trustworthy, factual data. If we make data about often misrepresented societal facts – such as crime, immigration or employment statistics – more readily available and more easily usable by automated tools, we can increase the speed, accuracy and scale of fact-checking.
As part of this project, the ODI will build on the excellent work done by the Office for National Statistics in the UK, and their counterparts in other countries, making statistical data available in machine-readable formats and under open licences.
Our goal will be to increase the availability and quality of the data being used to train automated fact-checking systems, but also to ensure that this increased interoperability creates further positive impact.
We will work as openly as possible in this work. If you are involved in similar initiatives, or would like to contribute to the project through workshops or pilots, please get in touch.