Combating AI deepfakes to safeguard the UK General Election
As Microsoft launches its Check Recheck Vote awareness campaign, Hugh Milward, Head of Corporate, External and Legal Affairs, outlines the measures the company is taking to combat the threat of deepfakes and misinformation in the run-up to the UK General Election on 4 July.
There are more elections taking place this year than any other year in history with over 40 countries and more than half the world’s population going to the polls.
On 4 July, the UK will play its role in this momentous year for democracy.
But deepfakes – persuasively realistic videos made to look and sound like real people – threaten to spread disinformation and undermine this democracy we value so highly.
This is why Microsoft has taken multiple steps to protect electoral processes as part of the company’s strategy to defend and advance democracy around the globe.
As a long-standing partner to government, civil society groups and news organisations in the UK, Microsoft is committed to bolstering the UK’s resilience against bad actors who use AI-based manipulation to disrupt democratic processes.
Our work in the UK is built on international agreements, such as the Tech Accord to Combat Deceptive Use of AI in 2024 Elections, signed in February 2024 at the Munich Security Conference with 26 other leading technology companies.
Agreements like these reinforce our own global initiatives, such as Microsoft’s Election Protection Commitments, announced in November 2023.
Protecting & empowering
Crucial to upholding electoral integrity is cooperation with elected representatives themselves.
To support candidates and parties in the UK, we provide our AccountGuard service free of charge to politicians and others involved in the democratic process.
This equips highly-targeted organisations and individuals with an extra layer of protection, benefiting political parties, think-tanks, journalists, non-profits and others with our experience and expertise.
As deepfakes and misinformation surge globally, the tools to identify and report them need to be resourced urgently. This is why we’ve engaged with political parties and the UK Government on raising awareness and resources for these essential defences.
We also have a webpage – Microsoft-2024 Elections – where a political candidate can report to us a concern about a deepfake of themselves.
Check. Recheck. Vote.
It is vital for the health of democracy that voters can trust what they see reported online about elections, particularly in the UK over the coming weeks.
At Microsoft, we believe that media literacy is essential to help people sort disinformation from verified fact and manage the risks posed by deepfakes.
This is why we’ve launched a campaign to encourage people to be vigilant to the risk of deepfake manipulation in the election, and to check, recheck and vote.
The Electoral Commission has also issued advice to voters on disinformation and to campaigners on ethical and responsible use of generative AI.
Partnerships & collaboration
We also work with media partners, such as NewsGuard, to provide tools to help people identify and counter misinformation they view online.
NewsGuard’s independent journalists rate online news websites for their journalistic integrity, giving a “nutrition label”-style rating for each. This service is freely available to all, and we urge everyone to make the most of it.
Microsoft is firmly committed to tech sector collaboration on this issue, supporting civil society groups to equip the public with the information it needs.
In addition, we are working with UK political parties and candidates to deliver support and training as they navigate the new world of AI.
And we provide security assistance against nation-state cyberattacks through our Campaign Success team and AccountGuard programs.
Working together
To protect the public from AI-driven threats to democracy, we need to strengthen our collective action and develop new mechanisms for international cooperation and coordination.
Microsoft is playing its part to ensure democracy is not undermined, and that technology for good effectively combats technology for bad.