Populism and political parties often function as synonyms. Parties tend to encash whatever is popular. They have often resorted to populist measures, from providing free electricity, bicycles, laptops, and bus rides to directly transferring cash to the accounts of certain sections while seeking votes for parliamentary or Assembly elections. And these have often worked to the advantage of one or the other party. Naturally, such promises tend to find space in election manifestos.
The big development of the times we live in is artificial intelligence (AI), which has the potential to reshape virtually every aspect of human life. From healthcare and education to manufacturing, logistics, and even entertainment, AI is revolutionising industries, improving efficiencies, and opening up new possibilities. Given the enormous potential, it is therefore crucial for governments to develop robust AI policies to ensure the responsible and ethical use of this technology while maximising its benefits for society.
While governments in power do talk about such policies, it’s surprising how AI has not become an intrinsic part of election manifestos of political parties, since it has the potential to impact what is seen as populist. Take the case of providing jobs to the needy where conventional wisdom suggests it’s best to automate processes and remove human interference to remove bias and corruption.
However, as historian Yuval Noah Harari has shown in his latest book, Nexus: A Brief History of Information Networks from the Stone Age to AI, computers and algorithms are riddled with biases.
Harari has illustrated the case of racism in commercial face-classification algorithms, as discovered by Massachusetts Institute of Technology professor Joy Buolamwini in 2017. She showed that the algorithms were very accurate in identifying white males, but extremely inaccurate in identifying black females. For example, the IBM algorithm erred only 0.3% of the time in identifying the gender of light-skinned males, but 34.7% of the time when trying to identify the gender of dark-skinned females.
The bias had crept into the algorithms because of the data sets on which it was trained. AI is nothing but finding patterns in data, and most databases comes with biases. The algorithms studied by Buolamwini were trained on data sets of tagged online photos — taken mainly from online news articles. Since white males dominate the news, the algorithms were excellent at identifying white men but bad at identifying black women.
Imagine what would happen if a data set of this kind is used to identify the intended beneficiaries of any government scheme. The scam could be worse than diversion of funds to a favoured set of related parties.
Let’s take another case related to jobs. It’s understood that AI will disrupt the market, with some old jobs disappearing and the emergence of new ones. But how does one identify which ones would disappear and which ones would emerge? Conventional wisdom suggests there would be demand for highly skilled workers but routine functions would face the axe. However, Harari offers examples that again are surprising.
It’s easier to automate playing chess than, say, dishwashing, he says. Until the 1990s, chess was often hailed as one of the prime achievements of the human intellect, while nobody thought dishwashing was particularly challenging. It turned out, however, that a computer can defeat the world chess champion far more easily than replace a kitchen porter. Harari says that automatic dishwashers have been around for decades, but even our most sophisticated robots still lack the intricate skills needed to pick up dirty dishes from the tables of a busy restaurant, place the delicate plates and glasses inside the automatic dishwasher, and take them out again.
Similarly, society may today put a premium on doctors and discount nurses, but the fact is that AI has the potential to automate the job of the former but not the latter. Most doctors gather medical data, offer a diagnosis, and provide treatment. These tasks are essentially pattern recognition, and spotting patterns in data is one thing that AI does better than humans. In contrast, AI is far from having the skills necessary to automate nursing tasks such as replacing bandages on an injured person or giving an injection to a crying child.
Harai says these examples don’t mean that dishwashing or nursing could never be automated, but they indicate that people who want a job in 2050 should perhaps invest in their motor and social skills as much as their intellect.
Since AI is a complex and highly technical field, many citizens may not fully understand its implications. Hence, political parties should lay out their AI strategies in manifestos to signal commitment to addressing these challenges and ensuring that AI benefits everyone. By including AI in their electoral discourse, political parties can not only enhance public awareness and trust but also ensure that the future of AI is shaped by thoughtful, inclusive, and forward-looking policies.
This provides an excellent opportunity to engage voters in a critical discussion about the direction of AI development, ensuring citizens have a voice in how this technology is used and regulated. Not doing so will blunt the promises they make to the electorate as translating them into reality may pose problems. Imagine what will happen if after coming to power, leaders start taking the view that chess champions or doctors need more freebies and cash transfers than dishwashers and nurses. The worst could be branding computers as racist or casteist.