AI and the destruction of the objective conditions of democracy
In 'Stopping Silicon Valley', Gay Marcus explains how the race to make more and more money is behind the recklessness of launching an imperfect Artificial Intelligence product on the market.

Instead of reading a veteran trained in the analog world eagerly hijacked by the digital revolution, I would go directly to read the naturalness and forcefulness with which Gary Marcus explains the potential for destruction of the objective conditions of democracy if we continue to ruminate alone or resignedly on the threats looming behind social media and the imperfect and failed development of generative AI. In fact, Marcus already considers the battle of social media lost and accepts the impotence of states and public authorities to enforce and enforce the rules valid for the analog world in the virtual world. He is on the screen of generative artificial intelligence and that's why he wrote the book at full speed, after more than 20 years of professional dedication to the digital sphere. He's a fan of it, of course, as we all are, because its benefits, its applications, the expansion of the world it fosters in a single hand are simply mind-blowing, and its discoveries and progress are literally incalculable.
But precisely because he's a total fan of the digital revolution, he can better than us gauge the undesirability of this revolution and the causes that explain a diagnosis he reiterates in many parts of the book: the dissemination and unrestricted and free access across the planet to generative AI platforms has been premature because no one is capable of correcting their defects, their falsehoods, their biases, their inaccuracies, their vulgarities, beyond occasional or very circumstantial patches. The race for money , the putrid race to make more and more money for companies that are already the richest in the history of humanity, is behind the reckless imprudence of having launched an imperfect product on the market and whose perfection no one knows right now how to approach.
To prevent the damage from continuing, to prevent it from accelerating uncontrollably, to prevent any worse evils than the already very serious ones that exist today, Marcus suggests putting the brakes on and redirecting the investigation.
That's why the second definitive grace of this book is to provide citizens with arguments, resources, and even tools to encourage a radical shift that will help ensure the reliability, trustworthiness, and efficiency of extensive language models, which are what everyone plays with, i.e., ChatGPT , Gemini, etc. To prevent the damage from continuing, to prevent it from accelerating uncontrollably, to prevent worse evils than the already extremely serious ones we have today, Marcus suggests hitting the brakes and redirecting research. Generative AI developments must be supervised, audited, co-financed, and guided by political powers, just as is the case with aviation or the stock market. Who would think of leaving the evolution of this revolution in the hands of the same companies that drive it and continue tolerating their complete opacity, their proven lies, their inconsequential abuses, their disdain for privacy, and their profiteering from everyone's usage data without paying anything in return?
Well, that nonsense is exactly what's been feeding the machine, without anyone being able to respond to it with any guarantee of effectiveness. The somewhat hippie, or naive, or simply progressive suggestion is that the driving force behind this change in operation will have to be citizen mobilization and resistance, and then political powers and governments will come after it if they want to continue winning elections. It's naive, of course, but I don't know if anyone out there has a better idea.
Jordi Gracia , El Pais Spain

Stop Silicon Valley
Gary Marcus
Translation by Marc Figueras. Shackleton Books, 2025. 256 pages. €22.90






