None of these were achieved by one person or even one organization.
The Challenges We Face Today
Every era is defined by the problems it tackles. At the beginning of the 20th century, harnessing the power of internal combustion and electricity shaped society. In the 1960s there was the space race. Since the turn of this century, we’ve learned how to decode the human genome and make machines intelligent.
None of these were achieved by one person or even one organization. In the case of electricity, Faraday and Maxwell established key principles in the early and mid 1800s. Edison, Westinghouse and Tesla came up with the first applications later in that century. Scores of people made contributions for decades after that.
The challenges we face today will be fundamentally different because they won’t be solved by humans alone, but through complex human-machine interactions. That will require a new division of labor in which the highest level skills won’t be things like the ability to retain information or manipulate numbers, but to connect and collaborate with other humans.
Making New Computing Architectures UsefulMaking New Computing Architectures Useful
Technology over the past century has been driven by a long succession of digital devices. First vacuum tubes, then transistors and finally microchips transformed electrical power into something approaching an intelligent control system for machines. That has been the key to the electronic and digital eras.
Yet today that smooth procession is coming to an end. Microchips are hitting their theoretical limits and will need to be replaced by new computing paradigms such as quantum computing and neuromorphic chips. The new technologies will not be digital, but will work fundamentally different than what we’re used to.
They will also have fundamentally different capabilities and will be applied in very different ways. Quantum computing, for example, will be able to simulate physical systems, which may revolutionize sciences like chemistry, materials research and biology. Neuromorphic chips may be thousands of times more energy efficient than conventional chips, opening up new possibilities for edge computing and intelligent materials.
There is still a lot of work to be done to make these technologies useful. To be commercially viable, not only do important applications need to be identified, but much like with classical computers, an entire generation of professionals will need to learn how to use them. That, in truth, may be the most significant hurdle.
Ethics for Ai and Genomics
Artificial intelligence, once the stuff of science fiction, has become an everyday technology. We speak into our devices as a matter of course and expect to get back coherent answers. In the near future, we will see autonomous cars and other vehicles regularly deliver products and eventually become an integral part of our transportation system.
This opens up a significant number of ethical dilemmas. If given a choice to protect a passenger or a pedestrian, which should be encoded into the software of a autonomous car? Who gets to decide which factors are encoded into systems that make decisions about our education, whether we get hired or if we go to jail? How will these systems be trained? We all worry about who’s educating our kids, who’s teaching our algorithms?
Powerful new genomics techniques like CRISPR open up further ethical dilemmas. What are the guidelines for editing human genes? What are the risks of a mutation inserted in one species jumping to another? Should we revive extinct species, Jurassic Park style? What are the potential consequences?
What’s striking about the moral and ethical issues of both artificial intelligence and genomics is that they have no precedent, save for science fiction. We are in totally uncharted territory. Nevertheless, it is imperative that we develop a consensus about what principles should be applied, in what contexts and for what purpose.