[ad_1]
As the general election draws nearer, researchers are cautioning that there is little time left for regulators to address the potential hazards of artificial intelligence (AI) impacting the democratic process.
Nevertheless, experts indicate that fears regarding deepfakes potentially altering election outcomes are likely exaggerated, as there is a lack of concrete evidence connecting an election result to AI manipulation.
According to a recent study by the Alan Turing Institute thinktank, there are signs that AI-generated content is already causing harm to democracy. Recently, an influx of supposed images and audio involving politicians has raised concerns.
For instance, Sadiq Khan, the Mayor of London, faced a scenario where deepfake audio purportedly featured him uttering contentious comments ahead of Armistice Day, which almost led to “serious disorder”.
Another incident involved AI-synthesized audio seemingly made by U.S. President Joe Biden, which appeared to discourage voting in the New Hampshire primaries earlier this year.
Sam Stockwell, a research associate at the Alan Turing Institute and principal author of the study, expressed concern that deepfake harassment against public figures might deter their participation in digital platforms.
Stockwell explains, “The difficulty in distinguishing AI-generated content from authentic material leads to a range of problems… It enables malicious entities to exploit the ambiguity by rejecting deepfake allegations as fake news, complicating fact-checking, and undermining democracy’s core tenets.”
The report urges media watchdog Ofcom and the Electoral Commission to collaborate on guidance and seek voluntary agreements with political parties for the ethical use of AI in campaign activities.
The publication also proposes that media outlets receive instruction on how to correctly report AI-fabricated false content and that voter information should include tips on recognizing AI-faked content and where to seek advice.
Dr. Christian Schroeder de Witt, specializing in AI and security research at the University of Oxford, adds that strategies for mitigating deepfake effects could include watermarking AI content and using provenance-based systems that link images and metadata in ways that reveal any tampering, since only original creators can alter provenance records.
Additional Reading:
AI specialists comment on unpreparedness for the largest election year
Warning: AI poses an increasing risk to the next general election
Meta uses artificial intelligence to combat election misinformation
Stockwell warns that “with the general election just weeks away, there is no clear policy or expectation set to prevent AI from generating false or misleading election information.”
“This is why it’s crucial for regulators to act promptly before time runs out,” he asserts.
According to Dr. Alexander Babuta, head of the Centre for Emerging Technology and Security at the Alan Turing Institute, “While it’s vital not to exaggerate the insecurity of our elections, especially without definitive proof of AI altering results, it’s equally important to seize this opportunity to fortify our electoral system against current threats.”
“Regulators can play a greater role in assisting the public to discern truth from deception and to protect the integrity of the democratic process.”
The Electoral Commission has addressed such concerns, with a spokesperson stating, “We are equipped to counteract misinformation regarding voting procedures and election management.
“Throughout the election period, we will distribute factual and reliable information, aiming to rectify any misleading claims concerning the conduct of elections.”
The spokesperson mentioned that the commission will establish an online hub to aid voters and provide access to pertinent resources on media literacy, while encouraging campaigners to act responsibly and transparently.
“The complex issues posed by AI are clear, yet overseeing campaign content would necessitate a novel legal infrastructure,” they highlighted.
“We will keep a close eye on the general election’s impact and are prepared to join discussions regarding how it affects voter trust and confidence moving forward.”
An Ofcom representative commented: “We’ll give thoughtful consideration to the recommendations within the report.”
[ad_2]
FAQ Section
What are the risks of AI related to elections?
AI poses risks to elections by enabling the creation of deepfakes, which can be used to spread disinformation or manipulate public opinion.
What measures can be taken to combat AI-generated disinformation?
Measures include issuing guidance on the fair use of AI in election campaigns, watermarking AI-generated content, deploying provenance-based solutions, and educating the public on how to spot AI-generated content.
Is there any evidence of AI-altering election results?
As of now, there is no definitive evidence that AI has altered the result of an election. However, the potential for such an incident is cause for concern among experts and regulators.
What are the responsibilities of the Electoral Commission and Ofcom?
The Electoral Commission focuses on combating false information relating to voting processes and election management, while Ofcom will consider the recommendations from the report regarding AI’s impact on elections.
Why is regulator action important ahead of elections?
Regulator action is crucial to establish clear guidelines and safeguards to prevent the misuse of AI in creating false or misleading electoral information, thereby protecting the integrity of the democratic process.
