SINGAPORE is seeking to ban deepfakes and other false digital depictions of election candidates, under a Bill introduced in Parliament on Monday (Sep 9).
The ban will apply to such content – including images, video and audio – published from the issuance of the Writ of Election till the close of polls on Polling Day.
The aim is to “protect Singaporeans from digitally manipulated content during elections”, said the Ministry of Digital Development and Information (MDDI) in a statement.
Deepfakes that misrepresent candidates, “if left unaddressed, can threaten the integrity of our electoral process”, said MDDI. “Voters must be able to make informed choices based on facts and not misinformation.”
Electoral candidates will be able to request that the Returning Officer reviews such content and issues corrective directions.
The Elections (Integrity of Online Advertising) (Amendment) Bill comes ahead of Singapore’s impending general election, which must be held by November 2025.
BT in your inbox
Start and end each day with the latest news stories and analyses delivered straight to your inbox.
What it covers
The new Bill seeks to amend Singapore’s Parliamentary Elections Act 1954 and the Presidential Elections Act 1991, to ban the publication of digitally generated or manipulated online election advertising that realistically depicts a candidate saying or doing something that he or she did not.
This includes deepfakes created by generative artificial intelligence (AI), as well as other forms of manipulation such as Photoshop editing and splicing or dubbing of audio or videos.
The ban will not apply to unrealistic animation, cartoons or memes; aesthetic filters or colour and lighting adjustments; nor campaign posters that are unrealistic or do not misrepresent candidates.
While the ban will not cover content published before the writ is issued, it will apply to boosting, reposting or sharing of such content after the writ. News published by authorised agencies is exempt, as well as private communication between individuals.
The proposed law applies to depictions of candidates. However, candidates are only confirmed on Nomination Day, which occurs between five days and a month after the writ is issued.
Therefore, from the issuance of the writ till Nomination Day, the ban instead applies to depictions of prospective candidates: individuals who have both paid their election deposit and consented to be publicly named.
To that end, for the first time, the Elections Department will publish the names of prospective candidates before Nomination Day.
Why the need
Singapore has seen a rise in deepfakes, especially of politicians, said an MDDI spokesperson. “While these have been used primarily for scams, it is not difficult to repurpose them for nefarious purposes during an election.”
While the government can tackle online falsehoods through the Protection from Online Falsehoods and Manipulation Act (Pofma), the new Bill provides “targeted levers” for misrepresentation of electoral candidates, said MDDI.
Pofma is used when the government knows the true facts, said the MDDI spokesperson. But for depictions of candidates, “it is much more difficult for the government to establish what an individual said or did not say, did or did not do”.
“Therefore, we do need the individual to come forward and say that this is a misrepresentation.”
When requesting a review, candidates may choose to make a declaration as to the truth of their claims. Making a false or misleading declaration will be illegal, and candidates who do so could be fined or lose their seat.
Content will be assessed by the Returning Officer and a team of civil servants. This is likely to include a technical assessment of whether the content is digitally manipulated, as well as an assessment of whether it is true, based mainly on the candidate’s declaration.
The Returning Officer can then issue corrective directions to individuals, social media services and Internet access service providers to take down the offending content or disable end-users’ access to it during the election period.
Social media services that fail to comply could be fined up to S$1 million upon conviction. For others, non-compliance may result in a fine of up to S$1,000 or imprisonment of up to 12 months, or both.
“The Bill will allow the Returning Officer to address deepfakes swiftly, by asking for their removal, given the fast-paced nature of information flow in an election context,” said the MDDI spokesperson.
While the Bill applies only during elections, an upcoming move tackles digitally manipulated content in general.
The Infocomm Media Development Authority (IMDA) will introduce a legally binding Code of Practice requiring designated social media services to implement measures to prevent and counter the abuse of digitally manipulated content on their services. In the coming months, IMDA will engage social media services to work out the details.
Other countries have already passed laws to tackle AI-generated misinformation, especially during elections.
South Korea has imposed a ban on deepfake political campaign videos within 90 days prior to an election, while Brazil has banned synthetic content that will harm or favour a candidate during elections.
The Bill will be debated at the next Parliament sitting.
