Videos and images have been widely shared with claims that they show African immigrants committing crimes. Others are claimed to show violence and destruction by or against migrants.  

In all the cases Africa Check investigated, the visuals were taken out of context and not linked to recent events.  

Visuals generated using artificial intelligence (AI) tools have been less common but are still a problem. One video sent to Africa Check’s WhatsApp line appears to show an Ethiopian man telling the media that he will not leave South Africa, despite recent threats.  

The video looks realistic, but there are signs it was made with AI. The speech is unnaturally slow and precise, which is not what we would expect in a live interview. The sound is also very clear, even though the interview takes place outside and a South African flag is blowing in the background. This suggests there was wind, which we would expect to affect the audio of the recording. At one point, two interviewers appear to ask the same question at the same time. Subtle signs like these are enough to question whether the video is real. 

Africa Check searched for the interview on the website and social media accounts of the South African Broadcasting Corporation, whose logo is visible on one of the microphones, but came up empty. This confirmed that the video was fake.   

Another image, supposedly captured during service delivery protests which escalated into looting that targeted migrants, shows a group of men carrying a coffin and a marble headstone past burning debris on a street. The flaws were also difficult to spot, but a closer look revealed the image was AI-generated. 

For recycled visuals, tools like reverse image search can help show where and when they were first shared. But this may not work with AI-generated content. While there are clues to look for, the best approach is often still to look beyond the content and consider the context. 

Think about the overall message or feeling of the content. Ask yourself what emotions it’s trying to trigger. Is it meant to make you angry, outraged or disgusted? This could be a warning sign.

If the event shown is important or newsworthy, the lack of any credible reporting is another warning sign.

Along with fake quotes from prominent people, fake documents are also being shared, with claims that they originate from government departments.

For example, this poster appears to be from South Africa’s department of home affairs. It says that “all illegal foreigners must leave South Africa by June 30th”. But the South African police said the poster was unauthorised. 

Another more official-looking “public notice” also appeared to come from the department of home affairs. It offers a reward for information leading to the location of “undocumented foreign nationals”. 

The notice may look convincing at first, but it does not match the department’s usual social media posts. The fake document uses a different font type and size. Its letterhead is also different. 

Like with fake quotes linked to public figures, documents impersonating government departments can mislead, especially when they mimic the format of real government notices. 

But the devil is often in the details. Check that the contact information matches the telephone numbers and email addresses on the department’s official website. Has the national coat of arms been included, and if so, is it authentic? 

Poor grammar is another red flag. Legitimate communication from a government department is usually carefully edited.

Disinformation about migration is also spreading through quotes falsely attributed to public officials. 

Sometimes the quotes are accurate but shared without important context, such as this clip, which social users claimed showed Democratic Alliance member Michael Sun complaining about migrants in South Africa. But it had been taken out of context. The full speech was actually about crime in Gauteng province.

Other quotes are completely made up. For example, there is no evidence that South Africa’s chief justice, Mandisa Maya, made an “official declaration” saying anyone who protests or promotes violence on 30 June “will be guilty of a criminal offence”. The date refers to a “deadline” set by anti-immigration groups for “undocumented” migrants from other African countries to leave South Africa.

In another example, posts claimed that Botswana’s president, Duma Boko, gave a speech that criticised the treatment of Botswanan citizens in South Africa. But there is no evidence that Boko said this. 

By falsely linking quotes to influential people, disinformers can manufacture outrage or panic. 

If a quote sounds shocking or extreme, ask whether the person is really likely to have said it. Compare it with how they usually speak. And if you’re still unsure, check the person’s official social media accounts or look for reports from credible sources.

WhatsApp voice notes are notoriously difficult to verify. Without visual clues, you need to listen carefully to what is being said. Has the speaker identified themselves or said which organisation they represent? Does the audio sound like it could be manipulated? 

Like with other forms of disinformation, context matters. Ask yourself why someone might share or create a fake voice note. Who could benefit from it?

Most importantly, if you are not sure whether a voice note is real, don't forward it. 

Take, for example, voice notes warning people that foreign-owned spaza shops are selling poisoned food to children. 

In one voice note, a woman repeats something she says she heard “from a Somalian Christian guy”. She claims that spaza shop owners held a meeting where they discussed poisoning all the food, drinks, cigarettes and chewing gum in their shops.

“They’re going to inject it with a poison because they said that they want to make sure that when the time comes, when the 30th of June comes, they want to make sure that they have already killed a lot of people just before they return back to their homes.”

The woman ends with a plea: “So please, please guys, please just make sure that you no longer buy anything … from any spaza shops.”

The voicenote plays on real fears. In recent years, there have been several cases of food-related poisoning linked to goods bought from spaza shops.

But the voicenote gives little evidence of a planned effort by foreign-owned shops. The woman does not say who she is or what area she is talking about. Her only source is an unnamed “Somalian Christian guy”. 

A search for similar warnings from trusted sources, such as the South African police or the department of health, does not support the claim. The emotional plea at the end is also a red flag. Urgent and alarming claims are often used to spread disinformation, as people are more likely to share the things they feel strongly about. 

So while we cannot confirm where the voice note came from, there are enough red flags to distrust it. At the very least, it should not be shared further. 

Loren Landau, research professor at Wits’s ACMS, told Africa Check that “as long as there has been public polling, South Africans have had decidedly negative attitudes towards migration”. 

He added that false narratives had circulated for decades but were made worse by the advent of affordable smartphones and algorithmic social media, which allowed messages to reach those most likely to act on the narratives. 

He said: 

What is more dangerous… is the power of false information to help transform those attitudes towards action.

While misinformation is always harmful and misleading, when it fosters fear and social division, it runs the risk of political mobilisation, marginalisation, and violence.  

Inflated numbers and narratives that essentialise social groups as threats help naturalise real-world action to ‘address’ those threats… the effects will be harmful and long-lasting.

This is the precipice on which we now sit.