WASHINGTON, Oct 30: The growing use of artificial intelligence to create misleading videos and fabricated political allegations has emerged as a major concern ahead of the November 3 US midterm elections, with researchers warning that deepfake campaigns could influence voters, damage candidates’ reputations and undermine confidence in democratic institutions.
Several Democratic candidates, including Sherrod Brown in Ohio and James Talarico in Texas, have become targets of misleading videos containing fabricated accusations. The material has raised concerns about the growing sophistication of digital disinformation campaigns and the difficulty of identifying manipulated content during an election campaign.
Researchers believe the videos are part of coordinated efforts intended to exploit political divisions and influence public opinion during the final days of campaigning. The incidents have also highlighted how artificial intelligence can be used to manufacture apparently credible news reports and personal allegations.
One video targeting Brown reportedly impersonated a television news broadcast and used photographs of his former intern, Haya Akbik, alongside images of other former staff members. The material attributed false allegations to individuals who had no involvement in the claims.
Akbik rejected the accusations, highlighting the personal consequences of politically motivated digital manipulation. The incident demonstrated how individuals who are not directly involved in electoral politics can become victims when their photographs, identities or personal information are used to create misleading content.
Another fabricated video targeting Talarico featured an alleged former student making accusations against the candidate. Investigations reportedly linked the content to a covert Russian influence operation, adding an international dimension to concerns about election-related disinformation.
Cybersecurity researchers identified similarities between the activity and Storm-1516, a Russian-linked disinformation network previously associated with campaigns targeting former US presidential candidates Kamala Harris and Tim Walz during the 2024 election.
The similarities included overlapping online accounts that had circulated earlier propaganda material and subsequently distributed the latest misleading videos.
Digital forensics expert Hany Farid assessed that both videos were likely generated or manipulated using artificial intelligence. The findings reinforced concerns that synthetic media is becoming an increasingly accessible instrument for individuals and organisations seeking to influence political discussions.
The videos did not necessarily require sophisticated production techniques to attract attention. Even relatively crude manipulations can become influential when distributed rapidly through social media platforms, particularly when users encounter them without additional context or verification.
The video targeting Talarico reportedly attracted more than 500,000 views before researchers exposed its fabricated origins. The figure illustrated how quickly misleading political content can reach a large audience during a sensitive electoral period.
The incidents have renewed questions about the responsibility of social media companies in controlling the circulation of manipulated content.
Platform X removed the original posts and suspended accounts associated with the material. However, copies continued to circulate, demonstrating the limitations of relying solely on content removal to contain disinformation.
Once a misleading video has been downloaded, shared or reposted, removing the original publication may not prevent its further distribution. Copies can appear on different accounts and platforms, making it difficult for moderators and fact-checkers to eliminate every version.
The problem is particularly serious during elections because false allegations can influence public perceptions before candidates have an opportunity to respond.
Political campaigns traditionally rely on speeches, advertisements, interviews and public appearances to communicate their positions. The growing availability of AI-generated audio and video has created another challenge, as fabricated material can imitate the appearance and presentation of genuine news coverage.
Technology that was once largely associated with specialised research and entertainment applications is now increasingly accessible through consumer software and online services.
Generative AI systems can produce realistic images, voices and videos from relatively simple instructions. While these capabilities have legitimate applications in education, entertainment, marketing and creative industries, they can also be misused to impersonate individuals or manufacture false statements.
The political impact of such material depends not only on its technical quality but also on the speed and scale of its distribution.
A fabricated recording may be shared in private messaging groups, circulated by anonymous accounts and amplified by users who believe it to be authentic. Even after a correction is issued, the original allegation may continue to influence public discussion.
Researchers have warned that the spread of synthetic media could also create a broader problem of public trust. When manipulated recordings become common, people may begin questioning genuine videos and authentic statements.
This creates what experts describe as the potential for a situation in which real evidence can be dismissed as artificial, while fabricated material is accepted as genuine.
The challenge therefore extends beyond detecting individual deepfakes. It involves maintaining confidence in digital information and ensuring that people have access to reliable methods of verifying what they see online.
Digital provenance has emerged as one possible approach. It involves establishing information about the origin, history and authenticity of digital content. Such systems can help users determine whether a photograph, video or audio recording has been created or modified through particular tools.
However, provenance mechanisms are not a complete solution because not all content carries verifiable credentials. Material can also be copied, edited or redistributed without retaining its original information.
AI-based detection systems are another area of development. These tools examine visual inconsistencies, audio patterns and other technical characteristics that may indicate manipulation.
Nevertheless, detection technology faces an ongoing challenge because generative systems continue to improve. Methods that identify one generation of deepfakes may become less effective as newer models produce more realistic material.
Independent forensic examination, reliable reporting and cross-verification are therefore expected to remain important components of the response.
The issue also raises questions about how electoral authorities should respond when misleading content emerges close to polling day.
Election agencies must protect the integrity of the voting process while respecting freedom of expression and avoiding unnecessary restrictions on political debate. Establishing whether a video is fabricated, identifying those responsible and responding quickly can require coordination among investigators, technology companies and independent researchers.
The US midterm elections provide an important test of how political institutions and digital platforms manage these risks.
The growing use of AI-generated content also has implications beyond the United States. Countries preparing for national and regional elections face similar challenges as social media becomes an increasingly important source of political information.
India, which has a large digital population and extensive use of online messaging platforms, also faces the broader challenge of preventing manipulated content from influencing public debate.
The risks are not limited to political candidates. Deepfake technology can be used to impersonate government officials, corporate executives, journalists and ordinary citizens. Fabricated statements may create confusion, damage reputations or be used in financial fraud.
As a result, cybersecurity experts increasingly view synthetic media manipulation as an issue involving national security, public safety and digital trust.
Technology companies are under pressure to improve detection systems, strengthen reporting mechanisms and make it easier for users to identify manipulated material. At the same time, governments are being encouraged to develop clear rules that address malicious uses without restricting legitimate AI innovation.
The latest incidents ahead of the US midterm elections underline the need for cooperation between technology platforms, researchers, regulators, election officials and news organisations.
For voters, the developments highlight the importance of checking the source of politically sensitive videos, looking for independent confirmation and avoiding the immediate sharing of material that has not been verified.
The growing challenge is that visual evidence can no longer be treated as automatically reliable. As artificial intelligence makes realistic digital manipulation easier, the ability to establish authenticity will become increasingly important to the functioning of democratic societies.