The Impact of Deepfakes on National Security and Public Confidence

The term “Artificial Intelligence” was coined by John McCarthy, a professor of computer science, in 1956. Since its commencement, artificial intelligence has started shaping lives by assisting in medical diagnosis, facilitating faster decision-making, digital assistance, grammar correction, automating repetition, etc. Nowadays it is even helping in entertainment and making presentations more interactive and appealing. It also helps create engaging content be it text, LinkedIn posts, micro-blogs, audio, video, or photographs. However, artificial intelligence has posed regulatory challenges and concerns by making the tasks of improving and updating already existing content quick and easy.

One such development is “deepfakes”, a form of synthetic media. Synthetic media refers to a digital content which is created, generated, modified, altered, distorted by using artificial intelligence. Synthetic media can be audio, audio-visual, visual. Recently, the “Deepfake Technology” has emerged as a potential threat in terms of communication, information technology, social harmony, economy and critical infrastructure.

Synthetic media generated using deep learning algorithms imitates humans, their voice, body language and behaviour. Deepfake technology has made it possible to fabricate highly accurate audio, visual, and audio-visual content. Deepfakes create lifelike and quite convincing fake audio, video, and photographs using deep learning algorithms. This can be done to secure old memories and to develop pedagogy which may include showing internal vital organs by a video instead of dissection. But some people use this for business or profits. They create content which is not educational but a news piece which looks real and convincing.Such content is being used for creating unrest among the masses.These people create a new content or manipulate the existing video or image in a way that makes it look authentic and persuasive. Such deepfakes shatter the confidence of the public in media and the news, spread disinformation, and alter public views or perceptions. The circulation of such synthetic media when found false, fabricated also shatters trust of the public in media, governmental institutions.

There are several reported instances which have shown wrong information regarding elections, diplomatic relations, important examination information.What once used to be a source of entertainment is now becoming cyber crime, a threat to national security. This article explores and examines the intricacies of deepfake technology, its impact, and its coping mechanism.

Understanding Deepfakes

Deepfake is a synthetic media, which means artificial production, modification, and edition using some software that looks original. For example, a video in which a deceased person is found singing last night. This can be done by editing some old photographs and effects to give this photograph movement or motion and using the “sticker” option to provide new clothes and accessories. Whoever watches and doesn’t know that the person shown died ten years ago will believe it to be true. This media is called deepfake, which has nothing to do with reality or originality.

Common Deepfake Creation Techniques

Deepfake technology leads to the generation of fake content which often looks original. They look original as they mimic bodily gestures, sounds, and facial expressions. They are created using “Deep learning Algorithms” and “ Generative Adversarial Networks”, GANs. In this mode, there are two algorithmsーa Generator, and a Discriminatorー to create and refine fake content.

The “Generator” creates the data/content, while the discriminator checks the initial content for its originality. This process is repeated unless there is negligible scope for refinement. This is called creating a deepfake, which is fake but refined several times to look original and appealing.

According to the Department of Homeland Security’s report, several tools are available on the market that generate the deepfake in the matter of a jiffy. They are AI dungeons and other AI/ML tools.

Legitimate use	Malicious use
Entertainment or beautification like de-aging, adding colours and sparkle, sounds, music, etc.	Creating fake news and disinformation to disturb mental and societal peace.
Educational purposes like showing science experiments and microorganisms, dissection of animal and human bodies, and explaining chemical reactions.	Fraudulent activities to deceive the public for fraud, and other scams.
Legitimate use leads to innovation in media and entertainment.	Malicious use is a threat to privacy, security, and peace in society.
Reconstructing missing data like blurred photos, restructuring videos for archives, or preserving memories.	Cheating by impersonation will become easier.

Examples of Deepfake

Deepfake video of Joaquin Oliver(2020)

In an unfortunate event of a mass shooting by a youth at a high school in Parkland, Florida, one of the students named Joaquin Oliver died in 2018. On his sixth death anniversary, his parents released a deepfake video with the voice of their son urging the US government to control the use of guns. This was an awareness video to raise awareness and vigilance regarding the controlled use of guns. However, the intention of the parents was in good spirits; its purpose was to bring out a campaign highlighting the potential misuse of guns in the hands of youth, but it raises questions about the severe misuse of the deepfakes as it can bring the dead person back to life. This video tells that he had died in that incident, but if this hadn’t been told, then anyone unaware of this mass shooting would not have understood that it is an awareness video, and may have spread misinformation regarding the narrative.

This shows the profound implications of deepfake technology. This technology can be used to create deepfake videos of any political figure or any celebrity to misguide the masses and create internal disturbances, ruin friendly relations with neighbouring countries, and image among foreign states and in the international community.

Singapore

In 2023, a promotional video of then-Prime Minister Lee Hsein Loong and then-Deputy Prime Minister Lawrence Wong was circulated, showing the senior ministers promoting a cryptocurrency scheme. Afterward, a few other deepfake videos were circulated in which the senior minister Lee Hsein Loong spoke about international relations. These videos were created by impersonation of face and voice. The Prime Minister said all these videos are dangerous and potentially harmful to the country and its interests.

Such videos were created to fool the public so that they deposit their money in such schemes and get cheated or robbed. These videos are also a threat to the economy of the country. This also shatters the trust of the public in political figures. Even if some broadcast real videos addressing the nation, they may not be taken in the future. Such videos will be taken as fake videos.

The Prime Minister’s video regarding cryptocurrency is commercially induced, while later videos on international relations are misleading and sow distrust. They weaken national cohesion.

Russia-Ukraine War

When Russia began its attack on Ukraine, some cyber attackers made a fake video of Ukrainian President Volodymyr Zelensky. In this video, it looked like he was telling Ukrainian soldiers to give up to Russian forces, but that was not true. This video was created using AI to make it look real, but it was fake. The video was shared on a news website and spread across the internet; however, it was removed even before the public could find out.

During the war, Russia has also been suspected of using cyberattacks to cause problems in Ukraine. These attacks messed up important things like phone networks, electricity, alarms for air raids, ATMs, and other payment systems. In a case, these attacks even caused a power outage during a missile attack.

Fake videos made with AI weren’t just aimed at Ukraine. Some videos targeted Ukraine’s friends in other countries. Another fake video tricked a former UK Foreign Secretary by pretending to be a call from a former Ukrainian president. In another example, a fake video showed a US spokesperson talking about Ukraine using American weapons for strikes in Russia. These fake videos show how powerful and dangerous AI-made lies can create conflict.

Philippines

A deepfake audio was circulated on social media in April 2024 which falsely showed Philippines President Marcos as instructing his military to respond to Chinese attacks in the South China Sea. The Government of the Philippines later denied this audio.

India

Rashmika Mandanna Deepfake case

Indian Actress Rashmika Mandanna’s deepfake video created by British Indian women named Zara Patel was viralized which the actress called “extremely scary” and she also requested for legislation for deepfakes in India. This video has face swapping done by AI. Her face was shown on someone else’s body.

Digitally altered video of Vice Admiral AN Pramod

This video showed Vice Admiral AN Pramod making false claims that Pakistan targeted Indian planes during operation Sindoor.

Impact of deepfake technology on national security

Deepfake technologies can be very distorting for the security of any nation. Audios, videos, and photographs can be altered and mislead the public and government officials. It can disturb internal peace by spreading wrong pieces of information and misguidance against the government, any public figure or regarding the launch of a welfare scheme for citizens. It can agitate citizens against the government by triggering their emotions. Let’s analyse the potential risks to national security through the threat of deepfakes.

Disinformation

According to the Global Risks Report, 2024, misinformation and disinformation are the biggest short-term risks, followed by ‘extreme weather events’. Many countries after the Covid-19 pandemic couldn’t recover thereby providing grounds for misinformation and disinformation to take hold. Countries suffered economically, socially, and politically. The cure for coronavirus led to research and its spread provided a breeding ground for the spread of wrong information through audio-video, and news channels. This information was about the end of the world, the threat to life, the side effects of Corona vaccine, the death rate of patients, and the negligence of hospitals which created mistrust and fear in the minds of the public. Malicious actors earned a good sum of money by resorting to these tactics. Misinformation coupled with artificial intelligence becomes deepfake. Manipulation of real-time scenarios by the use of video and photo editor software paved the way for such a nuisance. In place of showing challenges faced by the medical fraternity, crematoriums, pharmaceutical industries, and airport authorities false narratives and rumours were disseminated which led to tension, uncertainty, and division in the society.

Diplomatic communication

Deepfake technology can impact diplomatic communication in a derogatory manner to create internal conflicts in the country. The deepfake video of the Ukrainian president elaborates on this point. Where the troops were shown to have surrendered, this is a serious deepfake challenging the sovereignty, and integrity of Russia. It could have demoralised the military and spread wrong information regarding military operations. For example, if the military has to start an operation at midnight then this surrender video may lead them to believe that now it is not required and in turn getting attacked by the other side. It can be derogatory to the operational effectiveness of the troops.

Elections

Imagine a video in which a prospective election contestant supports the legalisation of Cannabis. Now it sparked a sense of disgust among the government and the masses. He was sent a notice from the highest court. Later on, it was found this was not him. It was a creation using AI. This is not a future threat, it’s an actual threat now.

Election Commissions of any country are not ready to regulate the deepfakes. Also, they are not equipped or skilled to monitor any deepfake. However, this is an offence under the election laws. Such things deceive the public. Such videos undermine the public interest in elections and the candidates.

Elections are essential elements of social cohesion. They are the cornerstone of the participation of the masses. They are the hallmark of democracy. They determine the future of power and add to the voice of the citizens to say something on any critical issue. The development of artificial intelligence, preferably generative AI has created new challenges for the integrity of elections. In 2023 and 2024 there were numerous such incidents of deepfake videos to discredit political opponents. This also has reduced the voter turnout on election day.

The impact of deepfakes on elections is mainly psychological or social. Deepfakes have been used for negative campaigning. They don’t directly motivate the user to vote for someone but to not vote for anyone or targeted contestants. Fear, mistrust, and insecurity are the most felt impact of deepfakes on elections worldwide.

Any contestant can be shown accepting gratification for something. A contestant may be shown promoting the trafficking of women for illicit purposes.

Deepfake Pornography

A deepfake pornography is basically a digital sexual abuse. It is a creation and fabrication of intimate images. In 2019, synthetic media expert, Henry Ajder and his team found that over 96% of the total deepfakes disseminated were porns. They added that the reason behind this is the easy applications for generating such deepfakes compared to a few years ago. Earlier people have to learn to use these tools by enrolling in courses but now everything has an app and easily downloadable and easy to learn, just some buttons.These deepfakes not just sexual abuse but also online abuse.

Deepfake pornography becomes a threat to national security when they are used to harass, bully, or extort social workers for their work, military personnel during any emergency, journalists covering a major accident, court case, a campaign, or any famous personality involved in any event where interest of the nation is involved.Such serious and sensitive videos can create psychological pressure, damage reputation of the person shown in those videos.

Deepfake Technology and Frauds

According to the SBI Deepfake scam guide, the growth of AI tools and its popularity among masses has changed the way of carrying out cyber frauds. Now they are deepfake scams, not just scams. AI can create fake audio, videos, and impersonate anyone as it looks authentic. Such fake content is being used to impersonate bank employees, relationship managers, to extract OTP, passwords and other sensitive information. These deepfake scams are more gruesome than traditional online scams.

Fraudsters collect photos, videos and other data from social media websites. Then using deepfake tools such malicious people create real looking audio, visual, or audio-visual content. Then they pretend to be someone from the bank of the targeted victim. They create fictitious identity cards to build rapport with the target, they contact the target through phone, messages, video calls. If the victim calls back then it starts with a tune which feels like he is being connected to a higher official. Sometimes they offer investment schemes, and also force the victim to part with the money in the pretext that the account will freeze or be deactivated.

Such instances lead to money loss, decline in the reputation of the bank. Victims often feel stress, anxiety, panic, threatened, unsafe after such scams. One must always verify unusual requests, or stalking through phone, messages in the guise of bank, customer care,etc.

Controlling People’s mind

Deepfake technology creates hyper-realistic audio, audio-visual,visual where the speaker or presenter looks real. The background music, filters, and other decorative tools, enhancement tools make it so impactful that the people shown in that content doing or saying those things which actually never happened. It includes face-swapping, face-reenactment, lip-syncing. Since people trust what they see or hear,deepfake can act on their emotions. It has the potential of influencing people’s behaviour or decision by manipulating their emotions. Repeated exposure to such fictional content which was created to make someone believe a situation which never occurred weakens independent thinking. If the public keeps watching edited clips, speeches, and false news either they will stop trusting the information or start believing the false information or disinformation. This way the public can be aroused with the feelings of hostility towards leaders, political figures, military forces, government itself. It may lead to internal wars, agitation, influence elections, division of society, researching the ability of people to verify the truth of the content so circulated.

The potential of deepfake videos to spark international conflicts

The cost of computing is going down year by year. Earlier accessing the internet facility was the dream of many people, but now everyone can access it. Internet coverage and low cost are the main factors in online information.

This does not even leave wartime affairs or peacetime affairs. Artificial Intelligence and Machine Learning have paved the way for opportunities for learning, expanding business, and disturbing peace. Deepfakes can be leveraged for wider purposes. Let’s identify potential mischiefs:

1) Military orders can be controlled or a military leader may be shown ordering anything against sovereignty and integrity, or friendly relations with other States.

2) The President or prime ministers may be shown defaming the other countries or their political leaders or public figures.

3) Any leader may be shown uttering racist comments that can badly disturb relations not with the country where that race resides and also tarnish the reputation in the international community.

4) A fake video of a popular daily informing the attack by the other country’s militants can ignite a sense of deep fear and security and lead to a stampede and killing of many lives.

Hence, deepfakes have the potential to destabilize nations and spoil friendly relations among each other.

Possible impact of deepfakes in India

India is a land of disparity. It is a deep blend of differences in geography, cultures, religions, languages, and even educational and economic opportunities. If in a deepfake video a political figure, a business person impersonated speaking against any specific strata of society then it may have serious repercussions including arson, killing, and even riots. This also kills the democratic character of our country.

Deepfakes can be used potentially for election propaganda. This may alter the public perception regarding any party or change their opinion about anything done before elections.

Deepfakes can also tarnish the image of the police by showing them harassing street hawkers or street children. In the name of sting operations, police may be shown receiving illegal gratification or torturing someone. It may be shown that the police are doing fake encounters to gather public attention. Also, deepfakes may be used to fabricate false evidence in a case that has taken media and public attention. Even if the video has been made to show something by way of a skit or recreating a scene it can be interpreted in a different manner that may be derogatory to the trial procedure. It may even yield to wrong acquittal or wrongful conviction. It may further lead to a loss of confidence in the Indian legal system in the minds of the citizens. This may also tarnish the image of India in international scenarios.

Detecting a Deepfake

We can attempt to find the fakeness of the videos by seeking the following cues:

1) Is the sound of the person shown in the video genuine? This is possible only when you have heard the voice of that person earlier.

2) Their eye movements should be like a human being.

3) Their lips should be in sync with the audio. If their lip sync doesn’t match with the sound it can be an artificial video created using a fake person who resembles the genuine person.

4) Facial expressions and skin flaws can also be a noticeable factor.

Challenges in Curbing the Deepfake

Whoever is creating deepfakes is well-versed with technology. They create deepfakes so that they evade the test of genuineness by deepfake detection technology.

Deepfake tools are freely available online. Anyone can use and leverage their talent to develop these deepfakes and easily disseminate them. Even if detected, it already has worked on the psychology of the viewer. Negative information creates a lasting impression over pleasing information.

Another challenge is low public awareness and digital literacy. Parents encourage their kids to be tech-savvy but kids are too young to understand ethical behaviour. They only keep sharing whatever deepfake falls in their log-in account. This way dissemination and money generation through user clicks keep increasing and deepfake creators stay motivated.

Deepfakes laws in India

The government in order to curb the menace of Deepfakes created by malicious people has updated The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 on February 10, 2026. These updated rules have inserted Section 2 (wa) which defines “synthetically generated information”, as any information out of “audio”, “visual”, “audio-visual” which is created,generated,modified, or altered artificially or algorithmically, using a computer resource. This is all done in a manner that such information appears real, authentic, or true. Such information depicts, or portrays any individual or event in a manner that the person looks indistinguishable from the natural person, and the event looks indistinguishable from the real-world event.

Now, the proviso protects the following legitimate uses:

Routine editing of content which is done under good faith which include formatting, enhancement, technical correction, colour adjustment, noise reduction, transcription, or compression which does not alter, or misrepresent the substance or the meaning of the underlying audio, visual or audio-visual information;
Routine creation of preparation, formatting, presentation or design of documents, presentations, portable document format (PDF) files, educational or training materials, research outputs, in good faith including the use of illustrative, hypothetical, draft, template-based or conceptual content, which is not false document or presentation;
Use of computer resources improving accessibility, clarity, quality, translation, description, searchability, or discoverability of the underlying audio, visual or audio-visual information in a way that no material part of the original content is generated, altered, or manipulated.

Challenges in Defining and Regulating Synthetic Media

Vagueness

In the definition of “Synthetically Generated Information” there are several terms which are not certain or distinct in their meanings. For example, “indistinguishable”,”good-faith editing”, “good-faith creation”. Let’s discuss one by one:

“Indistinguishable” means which appears to be the same or, one cannot tell the difference between the two. There is no standard test to ensure that the synthetic content has turned so level headed that it cannot be identified separately from a real person or real time event. For one person it can be indistinguishable and may not be distinguishable for the other. An ordinary person, a technically trained person, a forensic expert all have different standards to decide if the underlying audio, video, or audio-visual is indistinguishable from a natural person or real time event. What is now indistinguishable may be distinguishable tomorrow. Also, another challenge with this indistinguishable is that the same person can look different across different photos, videos, and sound differently in audios, audio-visuals because of health condition, weather condition, clothings, makeup, lighting, camera results, pixel quality, mood, facial expressions, nervousness, etc. This is a subjective standard. This may lead to original content being challenged as AI content. Absence of some criteria to determine indistinguishable may lead to inconsistent investigations, judicial pronouncements.

“Good faith editing” creates confusion as what is good faith while editing a audio, video, audio-visual content has no where discussed in the rules. This good faith is also subjective for different people. Cropping an image or changing the background may be routine good-faith editing for the designer but may be deceptive for some other. For example, changing the background so as to show that the couple was in Maldives but actually they were at Manali. In one view, it is only entertainment, but from the view of investigation or tracing anyone missing may be manipulative or deceptive.

In the same way, “good faith creation” is also vague in the case of content created for educational or research or demonstration purposes. Intention is not so easy to prove. It is manifested through the act, or omission. Simply speaking, by the outcome. “Good faith creation” tag may cause confusion or misinformation.

Fragmented Legal Framework

As of now, India has no dedicated legislation to curb “Deepfakes”. India has some existing laws or amended laws to manage the menace of deepfakes. India has the Information Technology Act, 2000 which regulates Section 66C, Section 66D, Section 67, Section 67A, Section 69A which provide some help in security from deepfakes but they do not address the present day challenges. Further, India has Bhartiya Nyaya Sanhita, 2023, a comprehensive criminal law which provides for punishments for various acts and omissions which are derogatory for public, peace, and morality. It defines several classes of offences. Broadly, it provides for punishment for cheating, impersonation, forgery, false information etc.

India has the Digital Personal Data Protection Act, 2023, where AI generated faces and voices without consent may be challenged, but it nowhere discusses or defines deepfake.

None of the law is a robust framework in curbing the deepfakes. We need a robust framework or strong legislation for deepfakes and protection of security of India from the creation of deepfakes.

Measures to curb deepfake

Artificial intelligence-generated synthetic media has posed significant risks to the nation by misguidance, allurement, arousing emotions, and false promises. Fake audio and videos can provide a breeding ground for the violation of laws, executive orders, and infringement of the privacy of citizens and various governmental departments. In order to maintain the trust of the citizenry it is important to take some measures to curb the dissemination of false information, the duping of online users. Following are the few measures that can be considered to detect, regulate, and limit deepfakes.

1) Policymaking and legislation

The governments are attempting to introduce a level of accountability. Europe has suggested the Artificial Intelligence Act and the United States has advanced the Executive Order on Artificial Intelligence. They have instilled trust in the AI value system.

2) Public awareness regarding deepfake technology

Public awareness regarding Artificial Intelligence and deepfake is crucial to help them differentiate between authentic and synthetic content. This awareness can be spread through school education, by the help of college students, by social media, and by newspapers. Awareness programmes should be backed by workshops and strategies to distinguish between human-generated and artificial intelligence-created content. This awareness programme may teach that if something is not backed by a source then it is fake and if the information is backed by the source then it is real. Training should include how this information is spread and how to verify its originality. What psychological tricks and mind manipulation are used by these deepfake disseminators?

3) Zero trust principle

Zero trust is a cyber-security framework that says: never trust, always verify”. It focuses on the role of mindfulness that can help equip users to combat attacks that are designed to exploit human emotions. It includes focus, clarity, concentration, self-regulation, dynamic allocation of concentration, and improved cognitive skills to minimise cyber threats. For example, if someone through a fake video attracts customers to a weight-loss programme then mindfulness will question: Is it trustworthy? Is it coming from an authentic source? Is it a real website? Who is the owner? One will not get swayed by fake promises of fake video. This is self-regulation and mindfulness. The digital age provides a platform where advertisers or malicious actors constantly compete for end-users’ attention, mindfulness can make a difference in cybersecurity outcomes which requires a split-second decision to emerge as a survivor or victim of a cyber threat.

4) Effective detection software

There is a need to develop effective detection softwares or other devices to detect deepfakes and decipher the veracity of the content that is flowing in social media. There should be such filters on social media platforms that help users decide the veracity of the audio, video, or images they are viewing. Deepfakes should be limited to the expression of creativity and entertainment or spreading knowledge not to misguide or agitate someone.

5) International partnerships and collaborations

Deepfake technology gains its strength through its dissemination across borders. International cooperation is necessary to curb its dissemination by sharing technology and application of laws of various countries in consonance. If one deepfake is illegal in a jurisdiction then other jurisdictions should not legalise it. International partnerships are needed for a common strategy and its execution. Their strategies in law making and policy making should be put into effect by keeping in mind the world wide web, not the territorial boundaries.

6) Information literacy

Information literacy is crucial for lifelong learners. Information literacy encourages users to question the credibility of the sources. Users will be prompted to check the source of the information. Users will not accept any information out of excitement or sense of trust. Users will check the information at multiple sources including web sources, research papers, books, journals etc. Additionally, they will learn to organise information and create new information based on existing knowledge like consolidation, research, organising data and applying this information to solve problems.

7) Media literacy

The ability to access and analyse media information is called media literacy. It includes identifying sponsored content, propaganda, awareness of benefits and risks involved. For example, which children programmes are inculcating education and which is motivating them for non-adherence to their schools. Media includes television programmes, audios, videos, photographs, video games. This is about developing capabilities regarding synthesis of electronic information, fact checking, source verification, checking its credibility, learning to use fact-checking tools and ethical media behaviour.

8) Critical thinking

The challenge to curb the menace of deepfake technology cannot be attained by implementing legislation and technological measures alone. People as users should also be equipped with inbuilt tools to gauze through the correctness of the version. Human beings are very powerful and intelligent among all living creatures. It is way more powerful than Artificial Intelligence. Everyone should be academically trained to use their critical thinking skills to navigate the corrosive impact of deepfakes.

Conclusion

Deepfake technology has emerged as a serious challenge worldwide. It breaks public trust, and weakens national security. Deepfake have reportedly contributed to online scams. Scams include cheating by impersonation, identity theft, and spreading fabricated information. All this is achieved by creating fake clips of social activists, public figures, celebrities from sports and media, journalists, military officers. Criminals are using advanced tools to commit offences cross-border and it becomes very difficult to locate these criminals or their originator as police and other investigating agencies have a long way to go to equip themselves with the advanced technology. Training the security personnel and other investigation officers also creates a drain on the government treasury. Such training also distracts them from the main goal of these law enforcement agencies.

As AI-generated content is highly-sophisticated, distinguishing authentic information from fabricated generative output has become very difficult for individuals, organisations, security experts, and governments alike. Deepfake technology is not geography conscious. Online information can be viralized in a few seconds. Society is witnessing a phase in which reality can be altered, distorted truths can be dismissed as false, and trust can be shattered. The planned dissemination of a single fabricated video can cause public dissatisfaction, financial losses, and irreparable harm to individuals and organizations.

Deepfake requires strict oversight and technological safeguards. Government, digital platforms, law enforcement agencies, engineers and cyber-security experts must collectively develop a protocol for detection of synthetic media and its dissemination.