“Technology will continue to advance, there is no way of stopping it. There will therefore be many opportunities, but also great challenges, such as converting the skills of people who may lose their jobs or establishing ethical rules at the right time. The truth is that writing laws before the technologies are available in the field, is very difficult. Unfortunately, it is the technology that must lead. Trying to regulate in advance, before things happen, is likely to delay progress.”
Just as the European Parliament is making moves to “regulate” artificial intelligence, Rafael Reif, President of the Massachusetts Institute of Technology (MIT), an example of global excellence in technological research, wants to narrow, without excluding, the need for regulation.
Born in Venezuela in 1950, and an electrical engineer by training, he speaks to the “Corriere della Sera” at Eni’s headquarters in Rome, where he has come to sign with the CEO of the energy giant, Claudio Descalzi, the renewal of a collaboration agreement for research that began in 2008.
More specifically, Reif argues that rules have to be established, but only “when it is possible to predict that a certain technological development will have dangerous consequences. Creating laws on the basis of assumptions, when something is not predictable, can prevent the advance of technology, without knowing for sure whether what you want to avoid might just happen anyway.” He clarifies his position with a number of examples. “Let’s start with CRISPR, a technique that is used to modify the genome. In this case, it is not difficult to imagine some possible effects from intervening on genes and so there is a certain sense in establishing rules and ensuring that a certain direction is not taken.”
The second example concerns the artificial intelligences. “We first need to distinguish between different branches of study,” Reif underlines. “One is machine learning, i.e. the construction of machines that learn tasks to help man. In this case the devices learn only what we decide to teach them: they could get very good and evolve, but will not be able to take universal control. Only if they learned so much as to become super powerful compared with our brains, then we might be concerned, but we are still very far from this stage. In this sense, there is a second branch of artificial intelligence research that aims to replicate the way that man thinks. But this will take a long time, certainly not the 20-25 years that some suggests.”
He explains when the warning threshold should kick in, and therefore regulation, through the real case of a drone, because “the artificial intelligence that we might worry about is that applied to moving machines, such as robots, cars or drones. Take a drone, for example, equipped with weapons that can open fire. They already exist, but for now they are not able to learn and do only what they are told. The next stage is machine learning: I could, for example, provide 20,000 photos of a person and tell the drone to identify and kill them. But even in this case, it would still be performing my order. The next step, the one we should worry about, is the moment when a drone can make choices for itself: it could, for example, identify another person, decide that he or she is equally dangerous and eliminate them. When we give them this capability it will be a problem. But it will still be man who decides and it is man that needs, in this case, yes, impose outright bans, because we can already predict what could happen.”
Forecasting through knowledge is, on the other hand, what MIT does. In 2015 the Institute launched a five-year plan “for action on climate change” that has as its primary objective “the progress of knowledge on global warming.” It also aims to establish some markers: “We have collected enough data to say that global warming exists, that we are polluting the atmosphere, that we have 25-30 years before the climate will become irreversibly unstable,” says Reif, “and therefore, this is the period of time we have to build a decarbonised planet.” The president-elect of the United States, Donald Trump, claims “not to believe in man-made climate change.” And last November, at the UN climate conference in Marrakech – before demonstrating a more open position the following week – there were rumours of he intended to withdraw from the Paris Accords, the historic agreement to reduce emissions and keep warming below 2 degrees compared to pre-industrial levels. “Do you want to know if we are worried? A bit. But Trump said many things during the campaign, let’s wait and see what he really will say and do in the White House. However, we remain firmly convinced,” Reif continues, “and based on scientific evidence, that climate change is real and that MIT can not ignore it. I really hope that the federal government will also consider it as important. But even if it decides that it is not a priority, we will carry on our work.”
“In fact, at MIT, we are convinced that not only to we have to train students but that our task is also to apply knowledge to solve the challenges of the planet.” Since May, the Institute’s communication has focused on the “MIT campaign for a better world”: a highly idealistic container in which we have brought together the main areas of our work. “This ranges, of course, from the climate to studies on diseases of the mind in old age, a growing issue, since we are all living longer,” explains Reif. “In this field. a month ago, a technique was discovered for stopping and reversing Alzheimer’s disease in mice, and will now be tested on humans. Raising our objectives is an integral part of the “MIT” method: The campus is home to people of 152 nationalities and when you are dealing with such high aims, other issues, such as racial, disappear. It’s not a question of diversity, but what each person brings. Society in general should be inspired by this kind of coexistence.”
Reif’s visionary outlook does not, however, exclude a practical sense. “While the first goal of the climate plan is knowledge, the second is to develop the technology to decarbonise and initiate the transition to other sources.” He tells the story of a professor who walked into his office with a light bulb and a cable in hand. He took a sheet of paper and light bulb came on. On the sheet was the first prototype of a small, transparent and flexible solar cell, “with which we will one day be able cover entire buildings.”
“The third objective of the plan,” the president continues, “is to involve other partners, from energy companies to government and departments. In order to have an impact on reality we need to go outside the academy, and there are two paths. Doing it alone, through startups, of which there are a great many around MIT, for example, those focusing on the storage of renewable energy. And then through strong partnerships with “visionary” companies, like Eni. “In fact, Eni is one of the founding members of MITEI (the MIT Energy Initiative), which is the Institute’s centre that is working to find low or zero emissions solutions and directly involves the energy giants (the founders also include BP, Shell, Saudi Aramco, ExxonMobil, with Eni currently the main industrial sponsor).
“For the MIT professors,” Reif underlines, “a brilliant idea and the applause of colleagues is not enough: they want to reach the market and have an impact on a global scale.”
Agreement with Eni
Partnership renewed to accelerate the transition to clean energy
The President of MIT, Rafael Reif, and the chief executive of Eni, Claudio Descalzi, met in Rome on Saturday to renew for four years the research partnership launched by the Institute and the energy company in 2008. The agreement (to which Eni will contribute $20 million) foresees the confirmation the Italian energy giant’s position as one of the founding members of MITEI (MIT Energy Initiative) and support for the three main research areas of MITEI’s “Low-carbon Energy Center” into technologies to combat climate change: solar power, energy storage and the capture, use and storage of CO2. “We are delighted to be able to continue our consolidated and successful relationship with MIT,” commented Descalzi, “and Eni remains strongly committed to the pursuit of a strategy for the transition towards cleaner energy and the reduction of CO2 emissions. We see our partnership with MIT as key to completing and integrating our efforts in research and development.” The results to date include solar cells that can be printed on a range of surfaces, from cloth to paper, luminescent materials for intelligent solar windows and wearable electronic devices to improve safety at work. Programs have also been launched for the capture and use of CO2, energy storage and the enhancement of natural gas.
This article was translated from Italian; the original can be found here.