Did AI Just Enter Kenya’s 2027 Election? What Anthropic’s Claude Investigation Reveals

AI propaganda in Kenya may no longer be science fiction.

As Kenya heads toward the 2027 General Election, a recent investigation by Anthropic, the company behind Claude AI, has raised new questions about how artificial intelligence could be used to influence political conversations online.

Anthropic has previously reported disrupting an AI-assisted influence operation that allegedly used Claude to help manage more than 100 political personas, generate content and coordinate social media engagement.

And Kenya was among the countries connected to the political narratives examined.

But let's be clear: this does not prove that the Kenyan government used Claude or that the government was behind the operation. Anthropic's findings point to an influence operation, not a confirmed government campaign.

So why should Kenyans care?

AI propaganda could look surprisingly human

When we think about AI election manipulation, we often imagine deepfake videos of politicians.

But the bigger threat could be much quieter.

Imagine hundreds of social media accounts appearing to be ordinary Kenyans. They post political opinions, comment on trending stories and argue with other users.

Individually, each account looks harmless.

Together, they could create the illusion that thousands of real people support a particular political narrative.

That's where generative AI changes the game.

AI can produce huge amounts of content, adapt messages for different audiences and help coordinate online activity at a scale that would previously require a large human team.

Kenya's 2027 election will be highly digital

Kenyan political conversations already live on X, TikTok, Facebook, YouTube and WhatsApp.

That creates opportunities for citizens to participate in democracy, but it also creates opportunities for misinformation, coordinated influence and AI-generated political content to spread rapidly.

The challenge won't simply be identifying fake videos.

It may be identifying fake people.

What does this mean for Kenyan voters?

AI isn't inherently political.

It can be used to educate voters, translate information, analyse data and fight misinformation.

But the same technology can also be used to manipulate public opinion.

As we approach 2027, Kenyans may need to ask more questions before sharing political content:

Who created this?

Is this account genuine?

Is this information independently verified?

Are these really hundreds of different people—or one operation using AI?

The Claude investigation is therefore more than a story about one AI company.

It is an early warning about how artificial intelligence could reshape political propaganda in Kenya.

The 2027 election may be Kenya's most digital election yet.

And the battle for public opinion may increasingly be fought not just by politicians and their supporters, but by humans working alongside machines.

The question isn't whether AI will influence politics.

The question is how much—and whether voters will know when it happens.

Pamoja AI | AI, technology and innovation in Africa