Future Politics. Algunos párrafos que señalé mientras lo leía.

A principios de año leí el libro Future Politics: Living Together in a World Transformed by Tech de Jamie Susskind y me gustó bastante. Con lo fácil que es encontrar libros que dicen todo lo que va a ir mal con la tecnología este libro trata de mostrarnos algunas ideas positivas (y también qué cosas pueden ir mal o están yendo mal). No me veo con fuerzas para comentarlo, pero sí que señalé algunos párrafos (sin intención de ser completista ni nada de eso) y voy a ponerlos por aquí.

Sobre dónde se realiza la computación (el trabajo informático) hoy en día y las consecuencias. La potencia de cálculo está disponible para cualquier dispositivo (no hace falta que sea especialmente pontente) y además las máquinas no aprenderán solo de sus ‘propias’ experiencias, sino que forman parte de un enjambre informacional:

A final point about machine learning: it used to be that the computing power that fuelled any particular system was physically present within the system in question. The most powerful digital devices literally contained the processors that made them run. The arrival of cloud computing in the last decade or so has meant that the computing power no longer needs to be located within the device itself: like Apple’s Siri it can be accessed over the internet. This has major implications for the integration of technology, as it means that small devices can draw on big computing resources (chapter two). But it’s also important for machine learning because it means that machines don’t need to ‘learn’ from their own separate experiences; they can learn from others’ too, so that every machine in a swarm or fleet adds to the collective ‘intelligence’ of the whole.

Los cuatro factores de la información: se recogen muchos más datos, gracias a la actividad social alrededor de todo lo que se puede digitalizar (y se digitaliza); el coste del almacenamiento se ha reducido cada dos años a la mitad durante los últimos cincuenta; la explosión en capacidad de cálculo nos ha dado la capacidad de procesar lo que almacenamos; finalmente, la información digital prácticamente no tiene coste de reproducción.

Four factors have made this possible. First, much more data is gathered, because an increasing amounts of social activity is undertaken by and through digital systems and platforms. Second, in the last fifty years the cost of digital storage has halved every two years or so, while increasing in density ‘50-million fold’. Third, the explosion in computational power has given us the ability to process what we store. Fourth, digital information has almost no marginal cost of reproduction— it can be replicated millions of times very cheaply. Together, these factors explain why the transition from a print-based information system to a digital one has yielded such an explosion of data.

A veces se usa de manera intercambiable código y algoritmos, cuando no son estrictamente lo mismo. La idea de algoritmo se refiere a la descripción de cualquier aproximación matemática al razonamiento, cálculo y manipulación de datos. Se trata de un conjunto de instrucciones para realizar una tarea. Pero no necesita estar codificado para que lo maneje una máquina. Cuando hablamos de tecnología digital el algoritmo es la fórmula y el código es la expresión del algoritmo en un lenguaje de programación.

Code and algorithms are often referred to interchangeably, but they’re not strictly the same thing.The word algorithm can be traced back to the ninth-century Persian mathematician Abd’Abdallah Muhammad ibn Mūsa ̄ Al-Khwar̄ izmī. The translation of his name, algorismus, came to describe any mathematical approach to reasoning, calculating, and manipulating data.8 Today, the word algorithm describes a set of instructions for performing a task or solving a problem. It need not be written in computer code. ... When we talk about digital technology, the algorithm is the formula and the code is the expression of that formula in programming language.

Sobre los protocolos en las actividades humanas. ¿Qué sucede si nuestro médico ignora las pruebas y se deja llevar por su experiencia o intuición y el resultado es dañino para nosotros?

After a while you are attended by a human doctor, Dr Smith, who inspects your leg and tells you that it needs to be operated upon immediately. Concerned by this response, you ask whether that course of action is consistent with the X-ray results. Dr Smith replies that he has not looked at the X-ray results; he does not need to. He’s an experienced physician and has seen this a hundred times before—the leg needs to be operated on immediately. The surgery takes place, and it turns out that Dr Smith was wrong: there was no need to operate after all. This would have been clearly visible from the X-ray results. Unfortunately, because of complications arising from the surgery, you suffer long-term damage to the limb. You sue the hospital. In court, unsurprisingly, the judge finds that Dr Smith acted negligently by not looking at the scans— not because the law contains a rule specifying that doctors must always look at X-ray results, but because the law lays down a standard that doctors must exercise reasonable care or they will be negligent. It was negligent, the court finds, to refuse even to look at the scans.

Tal vez si hubiera habido más tecnología, el Dr. hubiera debido seguir mejor los protocolos para poder avanzar.

Now imagine what might have happened in the digital lifeworld. Dr Smith would have been required to consult Robots MD and JD before operating. Not to do so would itself be negligent. In fact, it might not have been possible to register a surgical procedure on the system without having consulted Robots MD and JD first.When consulted, these AI systems would have warned Dr Smith that he ought to wait for the X-ray; operating without doing so would be likely to cause harm and be seen by a court as negligent.

Esto puede llevar a situaciones como las de no poder hacer nada que no haya previsto el diseñador del sistema, que nos proporciona un sistema que nos da mayor libertad para hacer cosas, pero que también la restringe.

Liberty and Private Power One of the curious traits of digital technology, as we’ve seen, is that it can enhance and restrict our freedom at the same time. It frees us to do things that we couldn’t do previously. But it restricts us according to the constraints of the code.

Intercambiamos tener que sufrir algo de totalitarismo a cambio de la comodidad y los resultados que podemos conseguir.

The legal scholar Tim Wu, referring to this example, observes that ‘consumers on the whole seem content to bear a little totalitarianism for convenience.’

La capacidad de procesamiento de datos tendrá otras consecuencias: cuando alguien consiga un trabajo, será común que éste sea medido, observado y evaluado por algoritmos. Esto significa que los algoritmos decidirán quién consigue su medio de vida.

Algorithms, in short, will decide whether millions of people access the most precious thing the market has to offer: a livelihood. Once a person has a job, it will become more common for their work itself to be measured, monitored, and assessed using algorithms. They’re already used to predict when employees

Pero dejar que los algoritmos tomen esas decisiones no es necesariamente malo. Es posible que algoritmos bien diseñados pudieran eliminar los sesgos y prejuicios de los humanos. podría, por ejemplo, ampliar la base de candidatos para un puesto, más allá de las universidades y centros habituales. Y lo mismo podría pasar con los créditos, seguros… Los algoritmos podrían convertirse en un mecanismo de redistribución más justo.

Using algorithms and data to make these decisions is not inherently a bad thing. On the contrary, it’s possible that carefully crafted algorithms could eliminate the biases and prejudices of human decision-makers.With regard to work, for instance, affirmative action algorithms could be used to broaden the pool of successful applicants from beyond the usual colleges and institutions. When it comes to loans, housing, and insurance, algorithms could be used to widen access for those who need or deserve it most. My point, at this stage, is simpler: code (embodying algorithms) is an increasingly important mechanism of distributive justice.

Pero esto no tiene por qué suceder por sí solo, es necesario prestarle la atención adecuada.

It demands close political attention.

También pueden afectar a los precios. Si alguien midiera cómo usamos los sistemas (incluyendo fuentes externas de información) podría aumentar notablemente sus beneficios. Esto podría llevar, por ejemplo, a precios para cada persona, esto es, cobrar a cada usuario justamente lo máximo que está dispuesto a pagar por un servicio, en ese preciso momento.

Research suggests that if Netflix took into account its customers’ online behaviour (5,000 variables, including how frequently they visit IMDB and Rotten Tomatoes) it could increase its profits by more than 12 per cent. An extreme end result would be ‘person-specific pricing’ whereby algorithms are used to charge customers the precise maximum they are prepared to pay at that moment.

El libro dice muchas más cosas y creo que ha valido la pena leerlo y aquí sólo hemos dejado unos párrafos que nos llamaron la atención.