‘AI re-energises capitalism, giving it a fresh language in which to frame its sordid business’

Tanuj Solanki’s new novel The Machine is Learning follows Saransh, an employee of a life insurance company, who is part of a special projects group that is devising an artificial intelligence-based system which, if successfully implemented, will leave 556 employees across the country disposable overnight.

Saransh finds himself in a moral and ethical conundrum on meeting the people whose jobs are on the line because of his project. Meanwhile, his girlfriend, a high-principled ex-journalist, provokes him to question the nature of his work, which could destroy lives in his bid to be seen as a “pioneer” for helping his company realise its ambitions.

The 34-year-old writer’s debut novel Neon Noon was a darkly absorbing account of a man trying to heal a broken heart by travelling to Thailand for a sex-fuelled escapade. His follow-up, Diwali in Muzaffarnagar, was a collection of short stories, largely concerning life in the suburbs, set in his hometown in Uttar Pradesh. The book won him the Sahitya Akademi Yuva Puraskar Award in 2019.

In an email interview with Scroll.in, Solanki unpacked his new novel, its thematic concerns ranging from capitalism to artificial intelligence, and how it is of a piece with his previous works. Excerpts from the interview.

Before we get to discussing The Machine is Learning, what do you make of the 21st century’s technological, social, cultural, and psychological obsession with machine learning and artificial intelligence? I understand that the deal with the company in your novel is to fire people using AI, but beyond that, what do you make of our broader fascination with AI?
I think AI excites us primarily because of the possibility of human-like or even superhuman intelligence, which is essentially the whole spectrum from producing a song or a novel to AI creating newer AI and the machines taking over and so on. But, frankly, I’m not very interested in all that. My interest is more in the bloodbath it can cause in, or for, good-old capitalism.

The deal with the company in my novel is not that it wants to use machine learning to fire people. That’s too out there. That can’t be. The company even denies it. I would argue that trying to know the deal with companies, as is the case with my fictional one, is like peeling an onion. That firing people can reduce costs is clear, but understanding the difference between the two is crucial, too.

Wrapping ugly outcomes in euphemisms becomes something else when those euphemisms themselves serve as final truths for people involved. Firing people can be equated to reducing costs. Reducing costs can be equated to tinkering with processes. Tinkering with processes can be equated to implementing new tech. Implementing new tech can be equated to doing unprecedented things. Doing unprecedented things can be equated to being pioneers.

Different people in the organisation hold different parts of this chain as their purpose, and personal success often depends on how deeply you believe your part of the truth. Of course, no one would say that firing people is equivalent to becoming pioneers. That’s just silly. But when the entire chain works well, some people get fired and some people feel like pioneers.

Apart from being a set of powerful technologies, AI and ML are also, to large corporations, new euphemism-making devices, which is also to say new truth-making devices. They re-energise good-old capitalism, give it a fresh language to frame its sordid business in.

The Machine is Learning (TMIL) is quite the page-turner, very different from Neon Noon, insofar as TMIL is plot-driven whereas Neon Noon is mostly one vignette after another. While the protagonists of both books seem to find no meaning in their jobs, Saransh from TMIL is not as prone to diving down existential abysses as T from Neon Noon, and checks himself every time he wanders too far away into his mind.
I have changed as a writer since the publication of Neon Noon in 2016. One of the changes is that I regard plot as more central to my creative project now. But rather than clear causal linkages, I think that sometimes a sequence in a plot can remain at the level of a causal suggestion: perhaps this happened because that happened, what do you think?

When in early 2018, the seed of TMIL was beginning to take root in my mind, I knew that a narrative premised on a technology that is, among other things, about figuring out pseudo-causal relations between variables has to be plotty. In the novel, Saransh’s actions and the mental states leading to them are not connected by clear lines, so to speak, and yet the overall picture you get is of understanding him quite well. In AI projects, too, step-by-step explainability is an acknowledged issue, and yet the final outcome usually is the most probable one.

It’s interesting that you mention the difference between the current novel and Neon Noon. Both T from Neon Noon and Saransh from TMIL work inside corporations, yes, but Saransh is far more “of and in the world” than T.

T is immediately articulate in his criticisms of capitalism because it is easy for him to be so – he is mentally absent at his job and loathes the time he has to spend in the workplace. Saransh is completely different. His grievances develop slowly and without much express articulation. He’s very present in his job and finds it to be very important.

Saransh cannot ruminate as much as T: he has trained himself not to, he doesn’t feel he has the time to, he has the wisdom to see some things as inevitable, he is not a poet or a writer (which T is), and so on. I would say that Saransh is more incorporated, someone you’re likelier to meet in person.

It’s great that you do not simplify the technological processes key to understanding your story.
I didn’t want to make the novel simple on the technological side, beyond a certain point. Part of the reason why new algo-tech can have terrifying consequences is because the lay person isn’t invested in understanding its basic structure, complexities. So a novel that makes all of it too simple is, in a way, lying.

Saransh’s colleague Mitesh, yuppie family man and ideal company man, isn’t big on dwelling on ethics, be it as part of his job or the larger world outside, and he also does not have much time or sympathy either for social justice. Meanwhile, Saransh’s girlfriend, ex-journalist Jyoti, who feels she is always carrying a flaming torch for the unprivileged is the exact opposite. Saransh spends a lot of time with both characters, and the combined effect of these personalities has a strong bearing on the decisions he takes late in the story. What are your thoughts on Mitesh and Jyoti?
The novel is about Mitesh and Jyoti, too. Jyoti, I believe, remains somewhat enigmatic to the reader because she’s somewhat enigmatic to Saransh, who is also the narrator. The novel, however, strives to convey that although she presents herself as having a consistent worldview, she’s still a person who’s figuring herself out, and is therefore liable to shift her position over time.

Mitesh, on the other hand, is consistent. He’s consistent in his failures of empathy, which, combined with his emphatic sense of righteousness, inevitably appear vulgar to Saransh and Jyoti. In fact, Mitesh is so consistent that at his worst he can seem a bit caricaturish. You can predict what he will say next, which is not the case with Jyoti, say. And yet there are more Miteshes in the world than there are Saranshes and Jyotis.

I found this funny, because I’ve seen this a lot: Jyoti texts and asks Saransh if he’s a supporter of Narendra Modi before deciding to meet for a date. Why do you think someone would ask this question?
The essential question is: “Are you a safe person?” But it’s rude to ask it directly, I guess. So there are surrogate questions – to gauge whether a potential date’s masculinity tends to be toxic or whether he’s capable of empathy or whether he’s an Islamophobe, and so on.

For Jyoti, specifically, the question is less a surrogate for a safety question and more a direct query concerning Saransh’s political views. She just wouldn’t date an out-and-out “Modi supporter”. I know many women like her, so that conversation was easy to imagine.

The long conversation between Saransh and his boss Unnikrishnan, where they clash over their ideas about pretty much everything, reminded me of the final exchange between Judge Holden and the Kid in Cormac McCarthy’s Blood Meridian, where the Judge berates his protégé for showing “clemency for the heathen”.
Now that you mention it, Unnikrishnan does seem to be a bit like the Judge, though I must say the Judge wasn’t on my mind when I wrote Unnikrishnan.

I guess with Unnikrishnan I wanted to create a character who has achieved enormous success in corporate structures, not only through running the rat race well (which is, perhaps, Mitesh’s destiny) but through a deep knowledge of its concomitant iniquities and a sort-of Faustian agreement with them. Unnikrishnan’s journey, how he came to make that agreement, is a different novel. In my novel, he is the smartest person who is also probably the most evil person.

Saransh’s conviction in his own views solidifies during that long conversation with Unnikrishnan. To see Unnikrishnan as the enemy is, in many ways, the final stage of Saransh’s education.

Most people who have read your short story collection Diwali in Muzaffarnagar will find the connection between the story in it called Reasonable Limits, and TMIL. Now, Reasonable Limits was closer to Neon Noon in terms of style. The character was allowed to drift down into deep, dark territory, so to speak, but here, Saransh, as you say, is a man “of and in the world”. Why did you make him that way?
People liked Reasonable Limits standalone when it was published in a newspaper, but its inclusion in the collection Diwali in Muzaffarnagar divided opinion. For some, it didn’t fit well inside a book of stories about young people from a small town and their filial duties and their middle-class striving.

Reasonable Limits differed from the other stories in its form – it’s a single sentence over six pages – and in how (people mention this) it is plotless, drifty, “a bit all over the place”, et cetera. It took me some time to realise that there was one more reason for the complaint: Reasonable Limits is the only story in the collection that is about Work.

And Work sticks out – because it is an island, because it carries within it the injunction of its separation from Remaining Life, and because narrative arts concern themselves more with Remaining life than Work. All the other stories in Diwali in Muzaffarnagar were about Remaining Life.

But, tell me, isn’t the workplace the site where we are likely to face our biggest ethical concerns?

If I can use Orhan Pamuk’s categories – Reasonable Limits is the naive response to a bunch of feelings and questions about Work, while TMIL is the sentimental response. “[E]ven if I were to extricate myself from this particular narrative, the elsewhere I would go to would in turn bind me in a new way, impose on me another damning mode of participation[.]” – Reasonable Limits states this frustration, like a poet would, like Neon Noon’s T would, but TMIL concerns itself with the process, as to how a 21st-century knowledge worker arrives at this defeated and defeating understanding, and whether this terminal point can be avoided, and whether any small good can be done along the way. There is hope in every fibre of TMIL.

I also hope that TMIL, with its deliberate mixing of Work and Remaining Life, is able to convince some readers to reread Reasonable Limits and see it as a vital part of Diwali in Muzaffarnagar. I know that sounds weird, as if I wrote the novel to save the short story, but that feeling was there, I won’t deny it.

Saransh reminded me of Siddhartha, the protagonist of Satyajit Ray’s film Pratidwandi. Both men are aware of something intrinsically wrong with the world around them, and this creates a simmering discomfort within them, but they are indecisive about scratching that itch until the very end when they kind of explode.
There are some parallels, yes. It is clear that Siddhartha in Pratidwandi is to be seen as wavering between the poles of “selling-your-(body-and-)soul-to-the-machine” and “revolution”. For Saransh, the poles are not that extreme. In narrative terms, both men are in proximity with people who practically represent the respective poles that are pulling them.

But there is a big difference between the two men, too. Siddhartha has no Work, only Remaining Life; he’s unemployed. His complaints about the system are therefore the complaints of an outsider. His struggle is of wanting to become an insider (of something). For Saransh, who starts with being happy in a well-paying job, the anguish is the anguish of an insider who keeps learning more and more about the insides. His struggle is of wanting to be an outsider (of something). In this sense they are, in fact, opposites.

Please talk me through how you worked out the end of the novel, and how it changed, if it did, over the course of writing the book?
When I started writing the novel, I wanted it to be a novel of rebellion and the good side to win in the end. But as I wrote the novel, I realised that the good side winning was a fantasy and that I’m not a good fantasy writer. How does an individual win when the enemy is the Gordian knot of capitalism? Perhaps a sword made from iron smelted in the fires of Ragnarok could help.

But my novel didn’t have the scope to include that sword. I say this despite the earlier claim that TMIL is a novel of hope. The space for hope was created when I lowered my ambitions with the novel and reminded myself of its bildungsroman aspect. Saransh’s formation was enough, I realised; his consciousness at the end is what can become hope for the reader.

Reframing TMIL as a bildungsroman also freed me from some the necessities of drama – catharsis, denouement, et al. In dramatic terms, it can be argued that TMIL ends pretty much where it began; yet, I would argue that everything has changed because (at least) one man has changed.

But the dramatic impulse never leaves us, does it? I must say that I am, as I believe some of the readers might be at the end of TMIL, invested in Saransh’s future. I think I might write a sequel, in which Saransh meets Jyoti in Delhi. Whether publishers are enthused by the idea is another matter altogether.

Most of your protagonists are sensitive to or eventually become aware of the social inequalities and institutional injustice around them. What would Saransh, who found himself deeply moved by the image of Alan Kurdi long before the events of the novel, feel or think about the ongoing pandemic and how it has worked out for the poorest among us?
The Saransh of TMIL’s beginning would be working from home and slowly coming around to accept Mitesh’s theory that working from home has increased everyone’s productivity because people don’t get into non-sequiturs during Zoom calls. He would also be privately agonising over the plight of migrant workers while watching news in the evenings.

The Saransh of TMIL’s end would also be working from home, but without paying heed to the non-sequiturs related to productivity. He would be using his time or money or network to help those in hunger. He would argue loudly and often end in tears. He might be under surveillance.