Fact or Fiction or AI

A journey down the rabbit hole
We are witnessing a double movement of sorts, with tech innovation driving unprecedented connectivity while mainstream society polarises to its limits. The never yielding influx of information and opinions has brought in the dawn of ‘Post-Truth’ politics. Factual data is less influential in shaping public opinion than appeals to emotion and personal belief. The ubiquitous concern around Artificial Intelligence (AI) and its ability to blur truth from falsity should hardly surprise.

As individuals, we are attracted to information that caters to our personal prejudice. When there is say a news item that feeds our belief system, we do little to explore its veracity. Seamless ease of sharing across platforms i.e. personal/group messaging through WhatsApp and social sharing through Facebook and Twitter eases chances of virality.

The impact of misinformation can rapidly cross the threshold of nuisance value. Recent false reports by Pakistani agencies on a supposed Chinese rocket attack in Sikkim could have easily escalated an already tense border stand-off. Fake imagery and videos inflamed communal tensions from Basirhat and Baduria in West Bengal to Muzaffarnagar in Uttar Pradesh. Eventual peace may be brokered but the resulting damage to the social fabric can’t be undone.

Fake news can be either crude lies OR a subtle separation of context from facts. AI is the newest tool to grease this manipulation. As per a recent report published in The Economist, Generative Adversarial Network (GAN), an AI-based algorithm can create a fake video by training itself from existing footage, reproducing visual attributes in any context. GANs do not fiddle or re-cut existing video clips but creates digital alter-egos who can say and do what you want them to do, thereby becoming very difficult to spot. A team of computer scientists at the University of Washington used similar technologies to create a convincing video of former US President Barack Obama saying things he hasn’t said. Fringe elements and miscreants deploying such videos on social networks can effortlessly cause disruption.

Their effectiveness for now is limited by the quantum and quality of available footage, and requirement of heavy computing power. However, with Moore’s Law doubling processing speeds every two years, it’s simply a question of when these tools become accessible to consumers.

PM Modi has given a fillip to technology via various initiatives. Almost everything, from governance to business, is grabbing the digital space. India has c.450 internet users i.e. covering 35% of its population. This low penetration coupled with trends towards cheaper devices & data implies digitisation would only rise. Without concomitant safeguards including online and social media literacy catching pace, fake news is bound to flourish.

The legal implications can be serious. As per the Information Technology Act (2000), electronic records (such as audio and video recordings) are admissible as evidence for jurisprudence. This admissibility is subject to provisions of the Indian Evidence Act (1872), requiring mandatory authentication by one of the certifying authorities e.g. a Network Administrator or a Systems Administrator “related to the recording device”. Without extensive Machine Learning-oriented expertise and tech, it is not very far-fetched to imagine situations where a certified authority can be deceived, thereby stymieing the legal process and jeopardising innocent lives.

Manipulative fake news can send jitters down the financial markets and wipe out significant shareholder value. Though RBI, SEBI and the insurance regulator have already issued guidelines, most measures are still in nascence. Governments’ tendency to trail behind tech may have more serious consequences going forward.

The solution to counter this menace is not easy to suggest. Debunking fake news, doctored videos and inflammatory content is essential, but equally important is to authenticate the sceptics. Dubious messages or posts from family members or friends have a tendency to quickly acquire misleading authenticity. “People are quicker to assume they are being lied to but less quick to assume people they agree with are lying, which is a dangerous tendency”, as per Will Moy, director of Full Fact, a UK based fact checking agency. Is an institutional protocol the answer? Governments of Singapore and Malaysia run fact checking websites to clarify inaccurate assertions. China has strong censorship laws. However, these countries are not vibrant democracies like India. Moreover, there is always a possibility of misuse by the protectors themselves, and so robust checks and balances are imperative. Alternately, we can take a leaf from The Academy of Ukrainian Press (AUP) which has successfully campaigned to build media and information literacy in their national education curriculum. Teenagers trained under AUP are educating their parents and grandparents on the skills needed to tell fact from fiction.

Government support is essential but it can only do so much. A few agencies in India are already working towards this issue but they operate on a non-coordinated case-by-case manner. This is where programmes enhancing awareness of civil society plays an important self-policing role. Only a grass-root apolitical approach can confront this threat, lest we allow fake news to violate citizens’ right to be informed.

By Jatin Bavishi, a Post-Graduate in Economics and freelance writer