The urgency of good AI
The urgency of good AI
Countering misinformation in Bangladesh is the need of the hour. AI could make it better or worse
Dr. Harisur Rahman & Sakir Mohammad
Publish : 13 Jun 2023, 03:06 AM Update : 13 Jun 2023, 03:06 AM
Countering misinformation in Bangladesh is the need of the hour. AI could make it better or worse
Dr. Harisur Rahman & Sakir Mohammad
Publish : 13 Jun 2023, 03:06 AM Update : 13 Jun 2023, 03:06 AM
Nowadays, most videos and images uploaded or shared by humans have an AI touch. In other words, we appear to not be purely human anymore.
Our existence is embedded by machines, be it mobile, laptop, iPad, or any other device or technology we depend on.
Without machines, we can hardly do anything. So, we are not human only; we are the blend of machine and organism, which Donna Haraway called a cyborg long back.
With the use of these new technologies and machines, democracy can be easily manipulated.
Candidates can throw darts at the opposition and lionize themselves without even using their brains. An AI can do that for them.
Before the mass access and progress of AI, it was only done by targeted TV ads, yellow journalism, and lobbying.
In the 2016 US Presidential election, Cambridge Analytica played a crucial role in manipulating public opinion by targeting particular demographics and using customized ads. It left no one confused about the efficacy of targeted ads.
Those kinds of manipulation were only possible by big companies with their big computers. But with mass access to AI technologies, anyone with a decent computer can create deepfake content, which Nina Schick called the infocalypse (a portmanteau of information and apocalypse).
And those contents can be easily disseminated to any group of people because of the infamous algorithm of Facebook and other social media.
We need to understand that in the age of synthetic media, a person's voice and image can be cloned and create video content that is hyperreal (more than real).
AI now generates individual humans without any references or imitation from the individual. Just imagine an AI-generated website, “This person does not exist,” where all the human images are fake but look real.
Common users or even media-educated persons cannot detect deepfake videos or audio without deepfake detectors such as the “detectfakes” of MIT.
For example, an AI-generated fake image regarding a fire in the Pentagon, the United States Department of Defense headquarters building, was covered heavily by the Indian media and of course, it surfaced on social media heavily. It was later clarified by the Pentagon that there was no such incident.
But what about the uproar that the fake news created? What can we do to counter these kinds of fake news? And what if this kind of fake content is created to manipulate elections and spread fake news about the opposition?
On top of fake news, people have cognitive biases. When they see news or fake news that is aligned with their beliefs, they will immediately believe it, even worse -- spread it, without verifying it.
So, what can we do to combat all these multimodal problems?
Interestingly enough, fact-checkers can be AI itself. Imagine AI can generate antibiotics for superbugs or can educate the masses to teach about the danger of misinformation and disinformation.
It can also parse content, language, source, and interest in the news and verify contents that are available online. It can also educate people on certain critical issues.
But do we have the AI to combat misinformation? Do we want to combat misinformation?
Perhaps, the developed country that created such an intelligent system has a combating mechanism. But what about the developing and underdeveloped countries? The virus (AI) is available for everyone but not the cure. It is the same as climate change. The developed countries are creating tornadoes and the developing countries are suffering from it.
The idea is simple, AI is not going anywhere. Also, we want to preserve our long-fought democracy.
In this equation, we need the so-called “Good AI” available for everyone. The government, the election commission, and other stakeholders of democracy should take into consideration the use of AI to counter misinformation and disinformation related to elections.
And, that will be the best possible solution to combat misinformation and safeguard our democracy. As the Bangladeshi general election of 2024 is at the door, this issue should be looked at with even more keen eyes.
Fact-checking training for communication professionals and social media influencers should be provided with the help of national and international stakeholders. Tech giants such as Microsoft, Google, Meta, and Twitter should be made accountable to safeguard our democracy.
Harisur Rahman is an associate professor at the Department of Political Science and Sociology at North South University. Sakir Mohammad is the Editor in Chief of NSU Journal of Student Research (NSU JSR), North South University.