What happens when the world’s largest platforms no longer send audiences to the Web sites that created the content? And what does that mean for African media, already operating on fragile margins?

Moses Kemibaro, founder and CEO of the Nairobi-based marketing agency DotSavvy, offered a clear, urgent perspective on these questions, as well as an African blueprint for how media can reclaim trust, relevance, and commercial value in the age of AI during the recent INMA Africa Audience Development Summit.

News and digital platforms

Kemibaro began by confronting a difficult truth: The long-standing relationship between news publishers and major digital platforms has collapsed. For over two decades, media houses supplied content to Google and social networks in exchange for referral traffic. But today, as he put it plainly, “this is fundamentally broken.” 

He referenced a recent story in which the CEO of People Inc. accused Google of being “an intentional bad actor” after the company’s Google-generated traffic fell from 65% to less than 30% in just three years. This dramatic shift, Kemibaro explained, coincided with the explosion of generative AI tools following the public release of ChatGPT in 2022.

The underlying issue is structural. Google now uses a single crawler for both traditional search and its AI-driven “overviews.” Instead of directing users to publishers’ Web sites, the platform increasingly surfaces information directly within its interface. 

“Your Web site, your digital platforms don’t actually get any of that traffic, ” Kemibaro said. 

News companies then face an impossible dilemma: “Do we block the crawler? … Or do we allow the crawler, which means giving away content for free to train a competitor AI?” 

It is no longer a strategic partnership. It is not, he emphasised, “a win-win scenario.”

This shift is particularly consequential for African publishers operating in resource-constrained environments. The re-centralisation of value, where platforms capture more traffic, data, and revenue, creates a fragile ecosystem in which many media houses cannot survive on legacy models. 

As Kemibaro put it, “the implication is that there’s a re-centralisation of value away from the content creators … to the platform owners.” And this reality is not approaching but has already arrived.

Where AI fits in

Yet the story is not solely one of threat. AI, Kemibaro argued, represents both a transformative opportunity and a serious vulnerability. The continent faces what he called a “duality” in its relationship with AI — “whether AI is a miracle or a minefield.” 

On one side of the spectrum, AI offers enormous potential for improving efficiency. Automation can reduce the burden of routine tasks, such as transcription, research, and editing, freeing up journalists for deeper investigative work. 

“Don’t give a human a robot’s job,” he said, emphasising AI should support, not replace, journalistic craft. In fields like editing, VFX, data creation, and analytics, AI tools can accelerate workflows and raise output quality. Industries like tourism already demonstrate this potential, with more than 86.7% of businesses using AI for marketing and content creation.

At the same time, AI adoption in African newsrooms remains uneven and often unregulated. A case study from Zambia revealed a concerning trend: “60% of journalists use AI tools like ChatGPT … 74% have received no formal training … and 85% of organisations have no AI policy in place.”

This ad-hoc, bottom-up adoption creates what Kemibaro called “a dangerous governance vacuum,” where journalists risk unintentionally publishing biased or inaccurate material, violating copyright, or leaking sensitive information to third-party models. 

The cost of poor governance is not merely editorial but also threatens the credibility and survival of entire institutions.

More alarming is the rise of data colonialism, a term Kemibaro used to describe how non-African platforms scrape African content to fuel global AI systems without compensation. “A form of resource extraction,” he said, where African media’s “valuable data and content” are harvested by companies that then build products with little relevance or representation for African audiences.

Because most models are trained on non-African datasets, AI-generated outputs about African places, cultures, and people often contain errors, stereotypes, or simplistic caricatures. These inaccuracies, in turn, feed into flawed marketing decisions and wasted advertising spend.

Compounding this is the rise of weaponised AI. Kemibaro described a media environment in which “generative AI is actively polluting the African information ecosystem,” undermining public discourse, destabilising political environments, and eroding institutional trust. 

He cited examples of deepfake videos boosting authoritarian regimes, AI-generated “journalists” designed to plant propaganda, and a seven-fold rise in biometric fraud across fintech and e-commerce. The results are catastrophic. 

“The first order effect is that the lie is believed … the second is that audiences begin to disbelieve authentic information … and the third order effect … is the complete erosion of media’s value proposition.”

An offensive strategy

But even in this volatile landscape, Kemibaro sees a path forward, one rooted in transforming how African media defines and monetises value.

Instead of playing defense against AI and platforms, African media must adopt an offensive strategy. News publishers can, and should, begin monetising their content through licensing models designed for the AI era. 

Kemibaro highlighted initiatives like Cloudflare’s Firewall for AI, which allows publishers to control and charge for content scraped by AI systems. “The old defensive strategy is failing,” he said, noting that bots increasingly ignore rules. 

Instead, new models like pay-per-crawl (where AI firms pay a fee for each page they access) and pay-per-query (where publishers earn royalties every time their content informs an AI-generated answer) offer viable paths to sustainability.

However, economic innovation must be paired with governance. Kemibaro called for the creation of a Pan-African AI Code of Conduct, emphasising it cannot be a simple import of Western frameworks.

“It must be grounded in African context and also principles,” he said. 

This framework would establish standards for transparency, disclosure, data privacy and sovereignty, bias mitigation, accountability, and human oversight — ensuring the continent’s adoption of AI strengthens rather than undermines public trust.

Central to his vision is the creation of an African Media AI Trust Mark, a certification that would identify media organisations, agencies, and technology vendors that comply with the ethical code. 

In a low-trust world saturated with synthetic content, Kemibaro argued, “verifiable trust becomes the ultimate premium asset.” 

A trust mark would allow publishers to differentiate themselves to advertisers, command higher ad rates, and establish a virtuous cycle in which strong governance and high-quality journalism reinforce each other. 

“Ethical behaviour is rewarded with better monetisation… which in turn funds higher-quality journalism,” Kemibaro said.

This shift also requires a new approach to audience measurement. Rather than focusing on impressions or reach, African media must ask a more important question: “How many real people saw this in a verified, safe environment?” 

The future, he argued, lies in human-only audience datasets, hybrid analytics, and rigorous detection of non-human or synthetic traffic. Verified audiences — not raw numbers — will restore advertiser confidence.

Kemibaro also urged collective action, stressing that Africa can no longer rely on fragmented experiments or isolated initiatives. The stakes are simply too high. 

“The time for isolated experimentation is over; the moment for collective, strategic action is now,” he said. “If you want to go fast, go alone. If you want to go far, go together.”

Banner art: Adobe Stock AREE.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .