By Jedidah Mumia
The digital landscape has shifted from static content to an era of intelligent co-creation. As we navigate this era, the success of any media strategy relies on three interconnected pillars: the ability to generate high-fidelity assets, the engineering precision to guide those models, and the data-driven precision to engage the right audience.
IAWRT Kenya has taken the obligation to train women in journalism on the use of these tools. Through the help of trainers like madam Josephine Karani, Lucy Kilalo and Grace Kerongo, working with AI tools has been a walk in the park.
1. AI generated Media
Madam Josephine Karani extensively talked about how AI-generated media has evolved from simple text-to-image tools into sophisticated “multimodal engines” capable of producing hyper-realistic video, immersive audio, and interactive 3D environments.
She touched on the core features and modernized tools like DALL.E, Midjourney and Canva media. From abstract compositions to photorealistic landscapes, Karani leverages AI to create visuals that challenge traditional artistic boundaries. These tools can also be used to brainstorm plotlines, character development, and even writing entire short stories or screenplays.
The ethical aspect of using AI previously had the “Deepfake Dilemma”. Currently, the focus has shifted toward provenance and transparency.
Madam Karani made sure to remind her audience on the importance of acknowledging and labeling one’s work. Industry standards now mandate metadata “watermarking” to distinguish synthetic media from reality.
As journalists, we need to take accountability. Our ethical frameworks prioritize consent, ensuring that the likeness and style of human creators are protected from unauthorized training.
2. AI Engineering: Frameworks for Precision
The black box of AI is being unlocked through Prompt Engineering. “Prompt engineering is all about clear communication, but this time with AI,” said Madam Lucy Kilalo. She added that when prompting the user has to be specific and brief to avoid confusion. “Keep your prompts focused on your goals and guide the conversation from general ideas to detailed points.” She emphasized.
To get high-quality and reliable outputs, professionals use structured frameworks such as;
- COSTAR framework: This industry gold standard ensures no detail is missed by defining Context, Objective, Style, Tone, Audience, and Response format.
- Chain-of-Verification (CoV): This is a recent technique where the AI is prompted to fact-check its own initial draft before presenting it, significantly reducing hallucinations.
- Few-Shot & Chain-of-Thought (Cot): By providing the model with a few examples (few-shot) or asking it to “think step-by-step”, one can guide AI through complex reasoning tasks especially for data analysis.
Madam Lucy Kilalo also touched on a few principles of effective prompting. She emphasized on assigning a role/ persona. This would put the model into a specific context, dramatically improving the relevance of its output. She also mentioned that the user should be specific and provide context, simply provide the background information it needs to get accurate answers.
3. AI for Audience Engagement
Engagement in the AI era is no longer about broadcasting to a crowd it is about anticipating the individual. “We are our audiences, so we need to know our audiences and become them for survival.” Said Madam Grace Kerongo, one of the trainers.
How does that happen? There is similar Web data that can be used to engage the audience. These include;
- the core Traffic Metrics (The ‘How Many’?). These metrics will tell you the total number of visits (sessions) to the website. Unique visitors will also be identified. These are distinct individuals who visited the site within a selected time frame.
- Engagement Metrics (The ‘How Loyal’?) These, metrics measure the “quality” of the audience. It measures the percentage of visitors who enter the site and leave after only one visit (the bounce rate). It also measures the context fir news and pages per visit.
- Marketing Channels (The ‘From Where’?). This section tells you which platforms driving the audience to the news sites. They include; users who type the URL directly into the browser, traffic from Facebook, X, WhatsApp e t c (socials), referrals which are traffic from other websites.
AI now monitors subtle behavioral sequences such as how long a user pauses on a specific image to predict intent and serve the next piece of content before the user even clicks.
AI also creates fluid segments based on real-time mood and interaction history, ensuring that a price-conscious user receives a discount code exactly when they show signs of hesitation.
As Madam Josephine Karani, Lucy Kilalo and Grace Kerongo have demonstrated in their respective fields, the future of media is a blend of creative vision and technical precision. By mastering the frameworks of engineering and the ethics of generation, creators can build an audience connection that is both technologically advanced and deeply human.
