maaxx.ca

Jakarta –

Today some of us maybe their timelines or FYP (For Your Page) social media have been overwhelmed with content showing President Joko Widodo singing a song Rungkad, Asmalibration, And so on. Even several other important characters also appeared singing various songs.

It turns out that this content is not original content, but the result of processing using artificial intelligence (AI), where a person’s voice can be processed and modified according to the lyrics and the melody of a particular song. For most people – if they are not informed that it is the result of the AI ​​processing – they can believe in the authenticity of the content. This phenomenon is also known as fake deepwhich is a combination of terms deep learningwhich is machine learning that mimics the way the human brain learns, i.e. mimics, and fake (FAKE).

However, some time ago, Geoffrey Hinton, also known as The Godfather of AI, reportedly resigned from Google in early May 2023 yesterday. As someone who played an important role in the birth of AI technology, the news of his resignation naturally surprised many audiences, especially regarding his motivations for resigning and leaving his field.

In several reports, Hinton expressed some of his concerns about AI, from the many abuses, including one for malicious activity, to his fear of AI development getting out of control. He also called on the public to be aware of this, especially the use of AI which is used to produce disinformation.

Of course, we can imagine how powerful it would be if AI was used for certain political interests, especially in elections. It’s not impossible, editing the voices of certain personalities using AI – which was originally only for entertainment – has been used for negative interests in politics.

This is actually not a simple fear hypothesis. For example, in the United States, President Joe Biden recently released a video containing a statement attacking transgender people in January 2023. In fact, the video is a clone of Joe Biden’s voice when he talks about the expedition tanks to Ukraine.

In response to this phenomenon, Facebook researchers recently claimed to be developing an AI capable of identifying deepfake, tracing the origin of content using reverse engineering techniques. Although they also admit that it is difficult to follow fake deep as an example of the edited video of President Jokowi singing various songs. Thus, an AI designed to fight AI requires time to develop.

Special consensus

Reflecting on the 2019 election, of course, to combat the hoax phenomenon, it’s not enough to be a curative effort like to come down content by filtering or based on reports. Moreover, before the next election in 2024, information about hoaxes should become more massive, coupled with the development of AI as it is now. So that, counters Hoaxes must also be countered by the AI ​​in order to balance the power of the AI ​​itself.

The government must be particularly concerned about this issue. For example, like Canada and Lithuania, which have formed special units or departments to fight against hoaxes. Both countries are also using AI to quickly verify and validate the veracity of information.

However, there are important notes that should be taken into account when the government uses AI to fight hoaxes. One of them is related to ethics. The government is required to have special standards and promote transparency to the public regarding the AI ​​system used to eradicate these hoaxes. A reference that can be used as a reference is Ethics Guidelines launched by UNESCO in 2021, which emphasizes the right to privacy, the protection of personal data and the transparency of the operation of AI.

The use of AI is also important to pay attention to transparent and inclusive labeling classifications. For example, to determine whether information is true or not, of course the measuring instrument used as a reference cannot be discriminatory and becomes a political tool to label certain parties.

Prevention efforts

Reflecting on the phenomenon of hoaxes which cannot be separated from electoral contestation, the efforts that must be made are certainly not enough for the curative efforts as described above. However, it is highly relevant if accompanied by preventive efforts through the development of digital civilization in Indonesian society.

It is also a big challenge, because currently the digital literacy level of the Indonesian society is only 62%. This number is the lowest in ASEAN where the average level of digital literacy in other countries is 70%.

Accelerated efforts are needed to achieve a higher level of digital literacy, both through formal education and through programs that target the community. Apart from this, an equal distribution of digital infrastructure is also important to ensure that internet access is distributed evenly throughout the country. In this way, access to information and digital skills can also be distributed fairly.

Accelerating the development of digital civilization is certainly not easy. Collaboration, seriousness and commitment from all parties are necessary for this to happen in various ways. With a good digital civilization, the public has a strong fortress to filter the information they consume to avoid the effects of hoaxes, especially during electoral contests.

Mohammad Iqbal Khatami researcher Independent Election Aware Committee (KISP)

See also “Preventing hoaxes during elections, Kominfo-Bareskrim strengthens the security of the digital space”:

(mmu/mmu)