Full Fact publishes new report on Facebook’s Third-Party Fact-Checking programme
Today we publish our second report on our experience of the Third-Party Fact-Checking programme run by Facebook.
We attempt in these reports to create further transparency around the Third-Party Fact-Checking programme, so that people can reach informed judgements and others can learn from it. We are independent from Facebook and not party to their internal discussions and choices so there are limits to the information we can report on.
In this report we give an overview of how the programme has evolved since July 2019, including examining where Facebook has responded to and/or implemented the recommendations we made in our first transparency report.
It’s our view that this is a valuable programme, in what has been a challenging 18 months. The ability to see and check content on Facebook and Instagram made a significant difference to our ability to tackle misinformation during the 2019 UK election and the 2020 Covid-19 pandemic.
Progress has been made by Facebook to address Full Fact’s previous recommendations, outlined in our 2019 report. But we continue to have concerns about transparency and scale, particularly the information that is made available to fact checking partners both within the product on the content being checked and on the impact that fact checks have on user behaviour.
On balance it is our view that the other internet companies should emulate the Third-Party Fact-Checking programme. In particular, from what we can tell, YouTube stands out as particularly being able to benefit from a similar programme.
This is echoed by conclusions from a separate research report, also published today by Full Fact, on the experiences of fact checkers from around the world in tackling online misinformation.
But a partnership such as the Third-Party Fact-Checking programme can only be one part of an effective response to misinformation and disinformation. Other decisions the internet companies make are critical and need scrutiny and oversight: from product design, to advertising standards, to rules for user behaviour. Most internet companies are trying to use AI to scale fact checking and none is doing so in a transparent way with independent assessment. This is a growing concern.
We also urge the UK government to play its part in responding to misinformation and disinformation. This week the government released its plans to tackle online harms, but it’s not yet clear if they go far enough. Internet companies should not make decisions on issues as fundamental as freedom of expression without proper democratic debate and oversight.
We make seven recommendations for Facebook and, where appropriate, other internet companies:
- Recommendation 1: In the queue, provide data points of number of shares over time, which can be displayed on a graph
- Recommendation 2: Invest in better claim matching to identify content
- Recommendation 3: Continue to collaborate with fact checkers when developing changes to the TPFC programme
- Recommendation 4: Continue to review whether users are given the right information when their content is fact checked
- Recommendation 5: Provide machine readable data to fact checking partners
- Recommendation 6: Share a register of emerging potentially harmful misinformation trends with governments and other relevant official bodies
- Recommendation 7: Implement greater transparency around the use of AI in claim matching
What is Facebook’s Third-Party Fact-Checking programme?
Since December 2016, Facebook has been working with fact checking organisations all over the world to help identify false content and try to reduce its reach.
Facebook now have a global network of fact checking partners from 80 organisations, working in 60 languages, all signatories to the International Fact-Checking Network code of principles.
Full Fact joined the programme in January 2019 as the first UK fact checker. In early 2020 Reuters also joined the programme as a UK fact checker.