In the age of free AI and deepfake software, democracy faces new threats
PETALING JAYA: Getting a free deepfake or artificial intelligence (AI) software is simply a matter of a few keystrokes, with several available on an open source platform called GitHub.
You can then superimpose the face of a person onto a video, and draft a dialogue to be read out by an AI-generated voice.
Although the process may not be so straightforward, it is easy enough for anyone with a little IT knowledge to create a fake video.
Such methods are now widely used by scam syndicates. But deepfakes are also being used in politics, ranging from analysing voting patterns and even to influencing public perception.
Social media analyst Assoc Prof Dr Sara Chinnasamy said AI can be used to spread disinformation to suppress voter turnout, with false messages created to discourage voters from casting their ballots.
ALSO READ: Teo: Govt working to make online environment safer
“AI can also be used to mimic the voice of a candidate claiming the results are rigged or to create fake news to disrupt vote counting, which could sabotage an election,” she said.
She said AI-generated deepfakes have been prevalent in US politics and similar methods have also been used to influence elections in Argentina and Slovakia.
While AI use is still not rampant in Malaysia, Sara said the easy availability of such digital tools may necessitate a response from the government.
“Many solutions go beyond regulating new technologies, and laws often lag behind even well-established technology,” she added.Sara warned that politically-tinged deepfake videos can spread like a wildfire on social media “because we have a population of over 34 million and nearly 31 million have access to the Internet”.
Political pundit Dr Tunku Mohar Tunku Mohd Mokhtar of the International Islamic University Malaysia (IIUM) believes AI is likely to influence political campaigns.
ALSO READ: Be careful with AI in elections, say political parties
“AI itself is a broad subject matter. In addition to videos and photos, writings can also be generated by AI,” he said.
However, the danger lies in the proliferation of fake narratives or images that can create anxiety among the public, he added.
He said apps such as Adobe Photoshop have also been used to circulate doctored content.
“It may also eventually create confusion and the erosion of trust in the media. We already see images of manufactured news items supposedly coming from the media. The public somehow takes it as real news.
“I think AI will be used both positively and negatively – positively to enhance the image of candidates and parties, and negatively as a tool to falsely discredit the opponents – in the next elections,” he added.
Asked if GE16 would be “dirty”, he said it was a possibility.
“Most dirty campaigning activities are done online or on social media,” Tunku Mohar added.
Computer science expert Prof Emeritus Datuk Tengku Mohd Tengku Sembok said the Internet has always been a political campaigning tool, and the latest developments in AI has made it easier and cheaper to produce propaganda, fake news and deepfakes.
“In GE16, creating AI-generated content will be cheaper than printing party flags, and more effective.
“Without proper initiatives to stop the irresponsible use of AI-generated content, this issue can grow big enough to consume the nation, affecting its multi-racial and multi-religious population,” he added.
To combat the abuse of AI, Tengku Mohd called for the strengthening of existing legislation against those spreading misinformation or using AI for malicious purposes.
He also suggested promoting digital literacy through public awareness campaigns and integrating media literacy into education, as well as developing AI detection tools to identify deepfakes and collaborating with social media platforms.
“Launch fact-checking initiatives to verify information, especially during elections,” he said.
