Deepfakes represent a powerful example of the capabilities and challenges of modern AI technology, but there are growing concerns about its use in upcoming general elections. However, the applications and implications of deepfakes are far greater than politics. We discuss.

 

With the ready accessibility of generative artificial intelligence (AI) platforms, such as ChatGPT, Gemini and DALL-E, the ordinary citizen has gotten a taste of what is possible with AI. However, for those who are considerably more proficient and have nefarious inclinations, AI is not being used to create benign content, such as articles, social media posts and images, but rather deepfakes.

Deepfakes have been around for at least the past five years, but the technology with which to do so has become more widely available, and so it seems to be coming onto the radar of some Caribbean policymakers. For example, in Jamaica, concerns were recently expressed about the use of deepfakes and AI-generated content to undermine the integrity of the democratic process (Source: Jamaica Observer).

Currently, Jamaica is in the throes of local government elections, which will be held on 26 February, and general/parliamentary elections are constitutionally due in 2025. Hence, to varying degrees, campaigning began in late 2023, and it is likely to continue until the general elections are held.

One of the lawmakers, Dr Christopher Tufton, was reported to be of the view that “the use of artificial intelligence in political campaigning is posing a threat to democracy, and he urged the legislature to discuss the issue with a view to establishing sanctions for people using the technology to create mischief.” He also opined that “regulation is urgently needed as the problem ‘[will only get more intense, either from the local government election but also a pending national election going forward over the next year and a half’” (Source: Jamaica Observer).

 

The impact on democracy should not be our only concern

Although AI-generated content can indeed be used for mischief during electoral processes as Dr Tufton has suggested, its use in the political space should not be the only source of concern. Essentially, deepfakes can be used to create highly realistic but entirely fabricated content making it challenging to discern truth from fiction, and so the potential applications are endless. For example, untrue content can be released to spread fear and create hysteria, and when compounded the virality of social media, can take on a life of its own that is difficult to reverse.

Further, and when focused on individuals, deepfakes can threaten personal privacy, as there is the capability to superimpose faces onto different bodies or insert individuals into compromising situations. Victims may find themselves falsely depicted in inappropriate or damaging content, leading to reputational harm or emotional distress, from which it may be difficult to recover.

Currently, and in the Caribbean region, laws or policies to help manage such situations – either to prevent them or to address the fallout – would be limited. In Jamaica, for example, the Office of the Director of Public Prosecutions has advised that there is no criminal liability for disseminating material which constitutes deepfake or AI-generated misrepresentation of facts. However, civil remedies could be pursued for the damage and only in very specific circumstances the actions may constitute a criminal offence. (Source: Jamaica Observer).

 

What can we do moving forward?

With at least eight Caribbean countries likely to have general elections between now and the end of 2025, the odds are the conversation and concerns about deepfakes will become more prominent. However, it is unclear what measures countries can – and will – take to address the issue.

So far, it appears that countries are inclined to strengthen the laws or otherwise try to introduce regulation. However, similar to efforts to regulate digital media, which were discussed in a recent podcast episode, the laws or structures put in place to manage deepfakes and AI-generated content could erode citizens’ rights. Additionally, the content may not necessarily be created locally, and the scope to enforce laws internationally may be limited. Hence, a deft hand is not only required in crafting such a framework but there also ought to be transparent and comprehensive engagement of the populace in its development.

Finally, the growing concerns about deepfakes also emphasise the need for greater digital literacy of citizens. Unfortunately, very few Caribbean governments are committing resources for such programmes or initiatives. Although improved digital literacy would not completely eradicate the problem, equipping people to better handle themselves in the digital space will be essential going forward, especially when it may not be possible, or advisable, to try to regulate every emerging development.

 

 

Image credit:  Freepik