Computers will always follow instructions. That may be the problem…

Many pessimistic scenarios about machines taking control of the world and harming humans are based on the idea that computers will eventually develop self-consciousness and define their own goals, incompatible with the goals of humanity. This is the basis of the argument of many science-fiction movies and books.

Many people believe, however, that this will not be the main problem. As reported in many news outlets, the University of California at Berkeley (my alma matter) has launched the Center for Human-Compatible Artificial Intelligence. The center will be headed by Stuart Russell, a famous expert in Artificial Intelligence (and  co-author, with Peter Norvig, of the most used textbook in the field, Artificial Intelligence: A Modern Approach). Russell has been a vocal advocate for incorporating human values into the design of AI, in order to avoid the pitfall that may come from AI systems running amok.

According to Stuart Russell, the issue is “that machines as we currently design them in fields like AI, robotics, control theory and operations research take the objectives that we humans give them very literally“. Therefore, they may approach tasks with an objective that is simply too literal. For instance, if instructed to solve the problem of “global warming”, a machine may decide that the most effective way is to wipe out the human race.


According to the UC Berkeley press release, the center is being launched with a grant of $5.5 million from the Open Philanthropy Project, with additional grants from the Leverhulme Trust and the Future of Life Institute.

The center will work on mechanisms to guarantee that the AI systems of the future will act, by design, in a way that is aligned with human values. According to Stuart Russell, “AI systems must remain under human control, with suitable constraints on behavior, despite capabilities that may eventually exceed our own. This means we need cast-iron formal proofs, not just good intentions.

Image credits: UC Berkeley. The image illustrates BRETT, the Berkeley Robot for the Elimination of Tedious Tasks, tieing a knot after watching others demonstrate it.

Is there life out there?

As reported in an article in the journal Nature, Proxima Centauri (pictured), the star nearest to our sun, has an Earth sized planet, orbiting the “Goldilocks” zone (not too hot, not too cold).

The recently discovered planet orbits the mother star in 11 days, an orbit much smaller and much closer to its sun than the orbit of the Earth. However, since Proxima Centauri is a red dwarf, it is much cooler than our sun, which makes this orbit to be just the right size. The planet, named Proxima Centauri b, weights between 1.3 and 3 times the Earth, which makes it likely that it may be a rocky planet. The distance to the star makes it possible that it may exhibit liquid water.


This combination of factors makes it the planet most likely to help us obtain additional information about the possible existence of life outside of Earth. Earth based instruments, such as the European Southern Observatory, ESO, an array of telescopes in the Atacama desert, in Chile, will be able to obtain additional information.

ESO was involved in the discovery of Proxima Centauri b, and likely to play an important role in the discovery of further information about this planet that, in astronomical terms, lies tantalising close to Earth, at “only” 4.2 light-years. Sending a spacecraft out to that planet may also be a possibility, albeit a very challenging one.

The challenges involved in obtaining further information about this planet are significant, but not unsurmountable, as the Economist reports. In a few years, we may have some better answers to Fermi’s famous question, “Where are they?”, referring to the possibility of extra-terrestrial life.


Uber to try self-driving cars, sooner than expected

Later this month, customers in downtown Pittsburgh should be able to call in a driverless Uber car. As reported by many news agencies, including CNN and Bloomberg, Uber will use Volvo XC90 sport-utility vehicles, equipped with sensors, radars, lasers and GPS receivers.


Although we have been expecting driverless cars to hit the streets some time soon, few have predicted that general usage autonomous vehicles would be available this year.

The partnership between Uber and Volvo makes the perspective of streets full of driverless cars less distant. Other companies, including Google, Ford and Tesla have their own plans for autonomous vehicles, but none of them has announced concrete steps towards making their cars available to the general public.

Uber cars will include a human supervisor, that will be in the vehicle at all times. Still, this development raises the prospect of job displacement in a massive scale, as CNN reports. Currently, Uber has 600,000 drivers in the US alone, and 1.5 million worldwide. However, as the technology for driverless cars improves, many more jobs than these ones that are at risk, as there are 3.5 million professional truck drivers in the US alone.

Reaching “longevity escape velocity”…

The concept that we may one day reach “longevity escape velocity“, a point in time when life expectancy increases by more than one year, every year, is not new. Many people believe that advances in medical and biological sciences will one day create the possibility that humans will live, if not forever, at least for millennia.

An interesting and very informative article in The Economist surveys some of the many ongoing efforts towards extending human longevity.


The “low tech” approach is based on the idea that calorie restriction (CR), the consistent ingestion of significantly less calories that what is normal, will significantly prolong life. Although the evidence is scant that CR is effective in normal humans, there exists some evidence that, under this regimen, other animals (and unicellular organisms) tend to live longer. The idea is that even a life extension of a few years may take you past the threshold where medical science may extend your life for centuries. So, a Pascal’s Wager makes sense: a few decades of sacrifice, in exchange for centuries of happy life.

More high-tech approaches include genetic manipulation and the development of special drugs that may delay ageing, such as metformim, resveratrol, or rapamycin. Clinical trials are at present very limited, because ageing is not considered a disease  and, as such, anti-ageing drugs cannot get regulatory approval. Self-experimentation seems to be very common in the field, though.

Interest in this type of research is likely to increase, as the population of developed countries ages, and the prospect of significant increase of life expectancy becomes more real. Believers in the singularity have one more incentive. After all, you only need to live enough to get to the singularity.

Explaining (away) consciousness?

Consciousness is one of the hardest to explain phenomena created by the human brain. We are familiar with the concept of what it means to be conscious. I am conscious and I admit that every other human being is also conscious. We become conscious when we wake up in the morning and remain conscious during waking hours, until we lose consciousness again when we go to sleep at night. There is an uninterrupted flow of consciousness that, with the exception of sleeping periods, connects who you are now with who you were many years ago.

Explaining exactly what consciousness is, however, is much more difficult. One of the best known, and popular, explanations was given by Descartes. Even though he was a materialistic, he balked when it came to consciousness, and proposed what is now known as Cartesian dualism, the idea that the mind and the brain are two different things. Descartes thought that the mind, the seat of conscience, has no physical substance while the body, controlled by the brain, is physical and follows the laws of physics

Descartes ideas imply a Cartesian theatre, a place where the brain exposes the input obtained by the senses, so that the mind (your inner I) can look at these inputs, make decisions, take actions, and feel emotions.


In what is probably one of the most comprehensive and convincing analyses of what consciousness is, Dennett pulls all the guns against the idea of the Cartesian Theather, and argues that consciousness can be explained by what he calls a “multiple drafts” model.

Instead of a Cartesian Theater, where conscious experience occurs, there are “various events of content-fixation occurring in various places at various times in the brain“. The brain is nothing more than a “bundle of semi-independent agencies“, created by evolution, that act mostly independently and in semi-automatic mode. Creating a consistent view, a serial history of the behaviors of these different agencies, is the role of consciousness. It misleads “us” into thinking that “we” are in charge while “we” are, mostly, reporters telling a story to ourselves and others.

His arguments, supported by extensive experimental and philosophical evidence, are convincing, well structured, and discussed at depth, with the help of Otto, a non-believer in the multiple drafts model. If Dennett does not fully explain the phenomenon of consciousness, he certainly does an excellent job at explaining it away. Definitely one book to read if you care about artificial intelligence, consciousness, and artificial minds.

Can Prisma and DeepArt make everyone an artist?

The popularity of Prisma, one of the hot summer apps (together with Pokemon Go), has caught everyone by surprise, including its creators.

Prisma uses deep learning algorithms to derive image processing methods that change your pictures in accordance with the style of a given artist. Other sites, like DeepArt and DeepDream, apply these methods based on machine learning techniques, such as the one described in this article, to process photos that you upload.

The following drawing of The Thinker was obtained applying Prisma to one of my travel pictures.


The following “painting” was obtained from one image of the tall ships in Lisbon, using DeepArt.


Applying the methods takes significant computer time, and is done by Prisma remote servers. These servers have, for a while, been unable to fully cope with the demand. Other sites, like DeepArt, also take significant time to process your request.

The results are, in many cases, surprising, obscuring the line between artistic merit and computerized image processing. Recently, Google raised a significant amount of money selling computer generated art.

For more examples of computer generated art, using Prisma and DeepArt, take a look at my deep art flickr album.