Artificial Intelligence (AI) is changing how we create, consume, and share information. Today, it is possible to generate realistic images, videos, audio recordings, and written messages in seconds. While these technologies offer opportunities for creativity, communication, and productivity, they are also making it easier to impersonate individuals, spread misinformation, and manipulate public conversations.

As Kenya approaches the 2027 general elections, the potential misuse of AI-generated content deserves particular attention. Fabricated political statements, manipulated videos, misleading images, and impersonation attempts could increasingly find their way onto social media timelines and messaging platforms.

For citizens, journalists, civic actors, and organisations working in the public interest, the challenge is no longer simply identifying false information. It is learning how to verify what we see, hear, and share.

What Are Deepfakes?

Deepfakes are images, videos, or audio recordings created or manipulated using AI to make someone appear to say or do something they never actually said or did.

For example, an AI-generated voice recording might imitate a colleague requesting confidential information or an urgent financial transfer. A fabricated video might falsely portray a public figure making controversial statements, while a manipulated photograph could be circulated as evidence of an event that never happened.

What makes deepfakes particularly concerning is their increasing realism. In the past, poorly manipulated images or videos often contained obvious inconsistencies. Today, AI-generated content can be convincing enough to mislead viewers who encounter it without context.

Seeing or hearing something online is no longer sufficient proof that it is authentic.

Why AI-Generated Deception Matters

The risks associated with AI-generated content extend beyond misinformation.

Impersonation and fraud: Attackers can imitate colleagues, organisational leaders, or trusted contacts to request money, passwords, confidential documents, or access to online accounts.

Misinformation and manipulation: Fabricated images, videos, and audio recordings can mislead audiences, distort public discussions, and circulate rapidly before their authenticity is questioned.

Reputational damage: Individuals and organisations can be falsely associated with statements, actions, or events, potentially undermining their credibility and public trust.

Phishing and social engineering: AI can help attackers create convincing messages that appear professional, personalised, and legitimate, making fraudulent requests more difficult to recognise.

For civic organisations, these risks are particularly significant. Public engagement, advocacy, and community mobilisation depend on access to reliable information and trusted communication channels.

Five Ways to Protect Yourself

1. Verify Before You Trust or Share

Before forwarding a surprising image, video, or audio recording, take a moment to investigate its source.

Who originally published it? When was it created? Is it being shared in the correct context? Have credible news organisations or independent fact-checkers verified the claim?

Be especially cautious with content designed to provoke anger, fear, excitement, or an immediate reaction. Emotional responses can encourage people to share information before checking whether it is accurate. A few minutes of verification can prevent misleading information from spreading further.

2. Confirm Unusual Requests Through Another Channel

Imagine receiving a voice message that sounds exactly like your supervisor, urgently requesting a financial transfer or confidential document.

Your first instinct might be to comply because the voice is familiar. However, AI-generated audio can imitate real people.

If a colleague or trusted contact makes an unusual request involving money, passwords, sensitive information, or account access, verify it through a separate, established communication channel.

A familiar voice, photograph, or writing style should not be treated as proof of identity.

3. Don’t Rely on Visual Clues Alone

Some deepfakes contain noticeable inconsistencies, such as unnatural facial movements, unusual lighting, or mismatched audio.

However, increasingly sophisticated AI-generated content may have no obvious visual or audio defects.

Instead of relying exclusively on appearance, focus on the source, context, and supporting evidence. Check whether the material has appeared elsewhere and whether independent sources confirm what it claims to show.

4. Learn to Use Digital Verification Tools

Several accessible tools can help investigate suspicious images and claims.

Google Lens allows users to search using images rather than text. It can help identify similar images, locate earlier appearances, and investigate whether a photograph is being presented in a misleading context.

Google Fact Check Explorer allows users to search for fact-checks published by participating organisations.

Image Whisperer provides another resource for examining images and investigating potentially misleading visual content.

These tools can support verification, but none should be treated as definitive proof that an image or video is authentic. Their findings should be considered alongside other evidence.

5. Protect Sensitive Information When Using AI

AI tools are increasingly used for research, writing, summarisation, data analysis, and everyday productivity. However, users should be mindful of the information they share with these platforms.

Avoid uploading confidential organisational documents, private communications, or personal information about colleagues, partners, and beneficiaries without appropriate authorisation and safeguards.

Before using an AI platform, review its privacy settings and data-handling practices. Consider whether the information being shared is necessary for the task.

The convenience of AI should not come at the expense of personal privacy or organisational security.

Digital Verification Is Everyone’s Responsibility

As AI-generated content becomes more widespread, digital literacy must extend beyond knowing how to use online platforms. It must include the ability to question information, investigate its origins, and recognise when additional verification is necessary.

This is especially important in the lead-up to Kenya’s 2027 elections, when public conversations, political messaging, and civic participation will increasingly take place across digital platforms.

Journalists, civic organisations, community leaders, and ordinary citizens all have a role in ensuring that misleading information is not amplified without scrutiny.

The most effective response to AI-generated deception is not necessarily becoming an expert at spotting deepfakes. It is developing consistent habits of verification.

Before you trust, share, or act on suspicious content, pause and verify.

That simple habit can help protect your personal information, organisational credibility, and the integrity of the public conversations we participate in.

Additional Resources

• Google Lens Tips (2025) — Practical guidance on using Google Lens to search images and investigate visual information.
• Image Whisperer — A resource for examining potentially misleading images.
• Google Fact Check Explorer — Search for existing fact-checks of circulating claims.