It is not everyday that Elon Musk endorses and partially funds a documentary. Needless to say, when that happens, anyone tracking the tech industry takes notice. In “Do you Trust Your Computer?“, Chris Paine brings together experts, journalists and CEOs from the tech industry and academia, to make a compelling point about the dangers of Superintelligence for humanity’s future. In this blog, I will review the documentary and offer some thoughts on how we respond to the issues raised by it.
In an age of misguided attention, I welcome any effort to raise awareness of the social impacts of AI. While AI has gained notoriety recently, there has been little thoughtful discussion of its impacts. I believe this documentary does exactly that and for that reason alone, I encourage everyone to watch it.
Surprisingly, the documentary did not uncover any new information. Most of the examples cited have been mentioned in other media discussing AI. The documentary contributes to the discussion not because of its content per say but because how it frames the issue of Superintelligence. Many of us have heard of singularity, the rise of killer AI, the death of privacy through Big Data and the dangers of automated weapons. Chris Paine’s genius was to bring those issues together to construct a cohesive argument that shows the plausability and the danger of the rise of superintelligence. The viewer comes away with greater clarity and awareness on the subject.
Compelling but Incomplete
In short, Paine argues that if we develop AI without proper safeguards, it could literally destroy us as a species. It wouldn’t do that intentionally but in the way to maximizing its goal. The example he gives is of how we humans have no qualms of removing an ant mound in the way to build a path. Superintelligent entities would look at us with the same regard we look at ants and therefore lack any human-centered ethical norms. Beyond that, he also touched on other topics like: the impending job elimination, Big Data’s impact in our lives and the danger of automated weapons. While the documentary was not overly alarmist it does challenge us to take these matter seriously and encouraging conversation at multiple levels of society.
In spite of its compelling argument, I found the treatment of the topic to be lacking in some aspects. For one, the film could have explored more how AI can lead to human flourishing and economic advancement. While at times it touched on the potential of AI, these bits were overshadowed by the parts that focused on its dangers. I wish they had discussed how, just like previous emerging technologies, AI will not only eliminate jobs but also create new industries and economic ecosystems. Surely its impact is bound to create winners and losers. However, to overlook its potential for job creation does a disservice to the goal of an honest dialogue about our AI future.
Moreover, the rise of artificial Superintelligence, though likely, it far from being a certainty. At one point, one of the experts talked about how we have become numb to the dangers of AI primarily because of all the Hollywood’s exhaustive exploitation of this theme. That was a great point, however, that skepticism may not be completely unfounded. AI hype happened before and so did an AI winter. In the early 60’s, many already predicted a take over of robots as AI technology had just entered the scene. It turned out that technical challenges and hardware limitations slowed AI development enough so that government and business leaders lost interest in it. This was the first AI winter from the mid-70s to the mid-90’s. This historical lesson is worth remembering because AI is not the only emerging technology competing for funding and attention at this moment.
Exposing The Subtle Impact of AI
I certainly hope that leaders in business and politics are heeding to Chris Paine’s warnings. My critique above does not diminish the importance of the threat posed by Superintelligence. However, most of us will not be involved in this decision process. We may be involved in choosing who will be at the table but not at the decision-making process directly. So, while this issue is very important, we as individual citizens will have little agency in setting the direction of Superintelligence development.
With that said, the documentary did a good job in discussing the more subtle impacts of AI in our daily lives. That to me, turned out to be the best contribution to the AI dialogue because it helped expose how many of us are unwilling participants in the process. Because AI lives and dies on data, data collection practices are fairly consequential to the future of its development. China is leaping ahead in the AI race primarily because of its government ability to collect personal data with little to no restrictions. More recently, the Facebook-Cambridge-Analytica scandal exposed how data collection done by large corporations can also be unethical and harmful to our democracy.
Both examples show that centralized data collection efforts are ripe for abuse. The most consequential act we can take in the development of AI is to be more selective on how and to who we give personal data to. Moreover, as consumers and citizens, we must ensure we are sharing in the benefits our data creates. This process of data democratization is the only way to keep effective controls on how data is collected and used. As data collection decentralizes, the risk of an intelligence monopoly decreases and the benefits of AI can be more equitably shared among humanity.
Moreover, it is time we start questioning the imperative of digitization. Should everything be tracked through electronic devices? Some aspects of our analog earth are not meant to be digitized and processed by machines. The challenges is to define these boundaries and ensure they are kept out of reach from intelligent machines. This is an important question to ask as we increasingly use our smart phones to record every aspect of our lives. In this environment, writing a journal by hand, having unrecorded face-to-face conversations and taking a technology sabbatical can all be effective acts of resistance.