submitted a legal brief in Mata versus Avianca that cited six court cases generated entirely by ChatGPT. None of the cases existed. When Schwartz asked ChatGPT to confirm the cases were real, the tool hallucinated again and assured him they were. Judge Kevin Castel sanctioned the attorneys and their firm, and noted plainly that technological incompetence is not a legal defence. Similar cases have since surfaced in Illinois and elsewhere, with lawyers fined for filing AI generated fabrications without checking a single citation.
Then there is the financial cost of not being able to tell real from fake. In early 2024, a finance employee at the Hong Kong office of the engineering firm Arup joined what he believed was a video call with the company ' s Chief Financial Officer and several colleagues. Every single person on that call, apart from the employee himself, was an AI generated deepfake. Convinced by the realistic audio and video, he authorised fifteen transfers totalling twenty five million dollars to five different bank accounts. By the time the fraud was discovered, the money was gone. Hong Kong police have since flagged this as part of a wider pattern of AI enabled corporate fraud.
These are not isolated glitches. Research from the SEC and Deloitte cited across recent industry reporting found that a significant share of companies globally have already suffered financial losses connected to deepfake attacks, and roughly a quarter of executives report their organisations have encountered at least one such incident.
The literacy gap inside companies
The uncomfortable truth is that most organisations are adopting AI tools far faster than they are building the human capacity to use them safely. A 2026 State of Data and AI Literacy report by DataCamp, surveying more than five hundred enterprise leaders across the United States and United Kingdom, found that seventy two percent now consider AI literacy essential for day-to-day work. Yet roughly sixty percent of those same leaders admit to serious capability gaps within their own organisations. Only around a third have a mature, organisation wide AI training programme.
Elsewhere, a Bright Horizons workforce survey found that forty two percent of employees expect their roles to change significantly because of AI within a year, yet only seventeen percent use AI tools
93 MAL74 / 26 ISSUE frequently today. Thirty four percent say they feel unprepared, and forty two percent say their employer simply expects them to teach themselves. The World Economic Forum projects that fifty nine percent of the global workforce will need reskilling by 2030, and that skills in AI exposed roles are changing sixty six percent faster than in other jobs.
Africa cannot afford to sit outside this conversation and assume it is a problem for wealthier economies. Digital inequality already runs deep, with a large majority of people in high income countries holding basic digital skills compared to a tiny fraction in low-income countries. As African corporates, government agencies, and media houses adopt AI tools for research, communication, and decision making, the same failure that embarrassed South Africa ' s government could easily repeat itself in a bank, a ministry, or a newsroom on this continent, unless deliberate literacy building happens first.
Beyond using AI, learning to detect It
AI literacy in the corporate space cannot stop at knowing how to prompt a chatbot. It must extend to detection. Professionals now need the ability to spot AI generated text, images, video, audio, and even chatbot responses that sound authoritative but are quietly wrong. This includes noticing the subtle tells in a synthetic voice, checking whether a quoted statistic actually traces back to a real source, and treating every AI produced first draft as something to be verified rather than trusted.
This is exactly where Media and Information Literacy becomes essential alongside AI literacy. UNESCO defines Media and Information Literacy, or MIL, as the set of competencies that allow people to access, evaluate, use, and create information and media content in a critical, ethical, and effective manner. Where AI literacy teaches you how a tool works, MIL teaches you how to question what that tool, or any source, hands you. Corporate professionals need both. Knowing how to generate a report with AI is only half the skill. Knowing how to interrogate that report before it reaches a client, a regulator, or a boardroom is the other, more important half.
Misinformation and Disinformation in the corporate space
Inside companies, this plays out in specific and often costly ways. Fabricated market research or hallucinated statistics can slip into strategy decks and investor presentations. Synthetic voice cloning is increasing ly used to impersonate executives in phone-based fraud, echoing the Arup case. AI generated fake product reviews and manipulated images can quietly damage a brand ' s reputation before anyone notices the pattern. Internal misinformation, such as an AI summarising a policy incorrectly, can spread through an organisation faster than anyone can correct it, particularlywhen staff trust the output simply because it sounds fluent.
Revealing these problems requires structure rather than luck. Companies benefit from mandatory verification steps before any AI produced content leaves the building, cross checking of citations and data points against primary sources, reverse searching of suspicious images, and clear internal reporting channels for anything that looks manipulated. Training staff to pause and verify unusual requests, particularly ones involving money transfers or sensitive data, remains one of the simplest and most effective defences against AI enabled fraud. None of this is complicated. What it requires is a workforce that has been deliberately taught to be sceptical in a useful way.
Literacy has changed, and so must corporate training
The lesson from South Africa ' s withdrawn AI policy, from Mata versus Avianca, and from Arup ' s twenty five million dollar deepfake loss is the same lesson repeating itself across industries. AI does not fail loudly. Yes, it fails quietly, convincingly, and often at the exact moment an organisation is not looking closely enough.
Corporate Africa has an opportunity to get ahead of this by treating AI literacy and Media and Information Literacy not as optional workshops, but as core professional competencies, the same way reading and writing once were. The businesses that invest in this now will be the ones still standing when the next confident, fluent, and completely wrong AI output lands on someone ' s desk.
Wafula Meshack is a Communication Specialist, Creative Writer and Associate Researcher at DIMLIS Africa and Kavod International. You can commune with him via on this or related matters via email at: Wafulameshack010 @ gmail. com.