TECH4GOOD
Americans will have their national elections in a couple of months. What makes this US election much more interesting than the previous ones is a phenomenon called deepfakes. It has added not just color but humor to the current campaign period. Some candidates are pulling out all the stops to win over voters, and what better way to do that than by creating deepfake materials?
Next month, the political circus will also be upon us when senatorial candidates start queuing at the Comelec offices to file their certificate of candidacies. This early, we are beginning to see the trimmings of election extravaganza marching hand in hand with the coming of the extended Christmas season in the Philippines. And I am afraid, just like in the US, we will also see widespread use of deepfakes aimed not only at getting our attention but also at spreading disinformation.
Deepfakes , AI-generated images, videos, or audio recordings, are a serious concern today. The rapid advancements in AI technologies have made these digital manipulations so realistic that even the most skilled experts find it challenging to distinguish them from genuine content. This difficulty in detection underscores the seriousness of the issue and the need for increased awareness and vigilance, especially during elections. However, there is another, more insidious form of deepfake that has been with us for so long: the deceptive politicians.
While not generated by AI, these individuals can be considered "human deepfakes," skillfully crafting a public persona far removed from their true intentions and beliefs. These politicians make many promises, becoming visible only during the campaign, manipulating information, misrepresenting their beliefs, believing that every voter can be bought, or engaging in deceptive tactics. They create a false reality, a carefully crafted persona designed to appeal to voters and advance their interests. This deception is often more subtle and difficult to detect than AI-generated deepfakes, but its consequences can be just as damaging.
One fundamental similarity between AI-generated and human deepfakes is their ability to manipulate public perception. Just as an AI deep fake can convincingly portray a person saying or doing something they never did, human deep fake politicians can present themselves in a way that is far from their true selves.
How do we fight deepfakes? Let's start with AI-generated deepfakes. The AI technologies used to create deepfakes are widely available today and are not limited to those who care about responsible use.
Deepfake generation methods are rapidly improving, making it increasingly difficult for detection models to keep up. As soon as a new detection method is developed, deepfake creators find ways to bypass it. Even highly accurate deepfake detectors can struggle to scale to the massive volumes of content being generated and shared online. Identifying 90 percent of deepfakes may sound good, but the remaining 10 percent can still have significant impact if widely distributed.
We need to check and recheck our sources. It is essential to be a critical information consumer and verify sources. The best way to stop the spread of a deepfake is to avoid sharing or commenting on it further. If a headline, video, image, or audio recording seems sensational, read the full story and verify its source before you trust, comment upon, or share it further.
Learning about deepfakes is essential in improving our media literacy and fact-checking skills. Many social media companies have built powerful protection features to detect and label AI-generated content and deepfakes so that users can understand their sources. If we see content we strongly suspect may be a deepfake, let us flag it to the platform.
How about deepfake politicians? Discerning voters may find them easier to detect. One would usually look for inconsistencies between public statements and their actions. If they have been making promises but consistently failing to follow through on them or are fond of vague rhetoric, it is a sign that they are deepfakes.
We also need to consider a politician's connections to special interest groups. If they consistently support policies that benefit these groups, it may be a sign that they prioritize their interests over the public good. We can listen to the feedback from voters who have interacted with them. If they report feeling misled or manipulated, it may be a sign that the politician is deepfake.
For years, we have seen how deepfake politicians' manipulations can have far-reaching consequences in the society. They erode trust in government, undermine democratic processes, and lead to policies that harm the public interest. While artificial intelligence may be advancing rapidly, the human capacity for deception remains a constant challenge. By recognizing the "human deepfakes" among us, we can work to protect our democracy and ensure that our leaders are genuinely working in the best interests of the people.
(The author is an executive member of the National Innovation Council, lead convener of the Alliance for Technology Innovators for the Nation (ATIN), vice president of the Analytics and AI Association of the Philippines, and vice president of UP System Information Technology Foundation. Email: [emailprotected])
