Dream Machine: AI and the Real World
Author: Appupen and Laurent Daudet
Publisher: Contxt
Price: Rs 599
Pages: 167
This book claims it is the first graphic novel on AI and it may well be. Even if it is not the first, it is undoubtedly the most detailed and thoughtful volume looking at both the promise as well as the dystopian future we are staring at if the large AI models being developed today fall into the wrong hands – that is, the hands of ambitious technology tycoons and global politicians aiming for world domination.
The graphic novel is authored by Indian born George Mathen, visual artist and musician, better known by his pseudonym Appupen. Well known for his dark humour and use of satire in his often-dark graphic novels, Appupen is the creator of the alternative world called Halhala, in which many of his works are based.
Laurent Daudet, the co-author, is a French physics professor and co-founder of LightOn, an AI startup working on GenAI for enterprises and founded with the stated ambition of pushing the boundaries of extreme-scale AI.
The collaboration has worked wonderfully though the final product is not for the faint hearted as well as those who are not already following the progress of AI.
The narration and even illustration styles change from time to time in the book. There is the straightforward story of a promising Gen AI startup founder Hugo Klein and his startup KLAI. Hugo is probably based on Daudet — certainly, it echoes his thoughts.
KLAI is about to enter into a deal with a tech giant called REAL.E, and if the deal is successful, it will catapult Hugo’s company into the big leagues. KLAI has recently achieved a breakthrough in training Large Language Models (LLMs) in Arabic and Hindi. REAL.E is interested in tying up with KLAI for a metaverse game it is building.
Being able to offer multiple languages to potential customers would dramatically expand REAL.E’s potential market size.
REAL.E wants KLAI to sign an exclusive contract —initially for Arabic and Hindi but later it wants KLAI to reserve all languages for it.
Hugo’s love interest is Anna— also a technologist but one who keeps Hugo grounded from the flights to fantasies to which he is prone.
Hugo has some faint misgivings about REAL.E but not major ones — until he meets Ayyo, an artist and storyteller from India, often creating comics built around conspiracy theories. Ayyo is probably Appupen’s alter ego.
The narrative shifts from the straightforward tale of a deal taking place in the AI tech arena to a superhero tale (SuperHugo) seamlessly.
There are three separate stories or storylines interwoven into the novel. The first is of course about KLAI and its REAL.E deal. The second tale is a masterful history of AI — short but without leaving out anything important. The third story is about the potential misuse of increasingly powerful AI models if they fall into the wrong hands. All the issues of Responsible AI and Ethical AI and the guardrails that need to be built into LLMs are explored with all the nuances.
REAL.E seems to be based on Meta (Facebook) loosely though it could be a composite of several big tech companies (Meta, Microsoft, Amazon, Google). The game it is building is a nod to 2022’s Metaverse mania which rapidly faded once OpenAI demonstrated ChatGPT and all tech companies started talking only about GenAI and LLMs and Foundation Models and the holy grail of Artificial General Intelligence (AGI), which is about machines becoming as good as humans in a range of general purpose tasks.
Hugo becomes increasingly worried that REAL.E’s game would suck out all knowledge from its players — and eventually replace them with virtual avatars who are even more powerful. It may well rob humans of all identity over time.
That isn’t the only danger. REAL.E also seems to be collaborating with a leader in Turkmenistan to create and operationalise a model that could ensure that no free or fair elections will ever be held and once someone comes to power, he or she can, with REALE’s help, remain in power forever.
In the corporate arena, as AI gets more powerful, it starts replacing humans in all functions with virtual personas that have sucked out all knowledge and made the earlier human functionaries who held these roles redundant.
Other dangers are also explored— from the fact that the new generation of AI models consume enormous resources (electricity, water, minerals and so on) from the earth – while their contribution to real-life problems such as global warming, education, poverty etc, has been minuscule so far.
This is an exceptional work looking at all the issues in AI in great detail — and illustrated perfectly. The story is well-told though the jumps between different narrations take a few minutes to get used to. Highly recommended for followers of AI— not for lay people though.
The reviewer is former editor or Business Today and Businessworld, and founder of Prosaic View, an editorial consultancy