Why would I ever need that?

This past Monday we talked about Internet voting. We didn’t talk much about the actual proposals for Internet voting and instead talked about how paper voting (or voting machines) compared with the aspirations for Internet voting. We also talked about how social media companies have experimented with their ability to influence voter turnout. As such, I thought I’d use this post to point you toward some actual Internet voting proposals. I’ll then end with a thought that came to me this morning while watching a news segment on selfies in the voting booth.

When I was an undergraduate, I met Andy Neff, a brilliant graduate student. Andy always impressed me as someone who could solve any problem, no matter how complicated. We lost track of each other during the late 1980s and early 1990s, and then we reconnected when I got interested in Internet voting schemes as part of Harvard’s Center for Research on Computation and Society. Andy had taken an interest in cryptographic techniques and their application to voting protocols, and he had become the Chief Technology Officer at a company called VoteHere. If there was anyone who could wrestle problems of Internet voting to the mat, I thought, it would be Andy.

While Andy and others made fantastic progress, even he will tell you that the world was not ready for Internet voting in the year 2005. To his credit, his work was one of several efforts that produced provably superior systems to the Direct Recording Electronic (DRE) voting machines that were being used in many polling places around the U.S. Unfortunately, he couldn’t solve every problem that could occur and any failure mode was enough to keep the world from moving to something that looked very different than the paper system. An excellent paper describing the issues that remained was written by David Wagner and his colleagues at U.C. Berkeley in 2005.

VoteHere doesn’t exist anymore, but if you want to dive into a system for verifiable online elections, I suggest you investigate Helios. This is a very popular voting system, but do note the answer to the last question in their FAQ.

This leaves us with the news segment about people taking selfies in voting booths that I saw on CBS This Morning. In some states, like Tennessee where singer Justin Timberlake took and posted a selfie of himself voting, it is illegal to take pictures in the voting booth. This law enforces the rule that, in most states, elections are held by secret ballot. Ballots are anonymous and by marking your ballot in secret, it becomes much harder for someone to intimidate you and influence your vote, or for someone to buy your vote. If you vote in secret and no ballot carries your name, how can someone who is trying to intimidate you or buy your vote know how you voted? Obviously, if you take a selfie of yourself with your completed ballot, this completely destroys the secret nature of the election.

I was further amazed when Charlie, Gayle, and Norah announced after the segment was complete that they didn’t see why it wasn’t perfectly fine to take a selfie in the voting booth. Have we gotten so far beyond the days of blatant voter intimidation and vote buying that no one feels that they need to vote in secret?

Let’s assume, for a moment, that the answer to this question is yes. How might we build a new voting system that takes advantage of the selfie movement? What if we all just took selfies of our completed ballots and submitted those “votes” to a server overseen by observers from both parties? They verified the count and could spot check the vote recorded for each of us was the vote we posted on our social media page. I’ve spent only a few minutes thinking about how you might do this, and so I’m sure I’ve forgotten some attack vectors. Still, what rules do we follow today for paper voting that no one seems to question why we continue to follow them? To be clear, I’m not saying every vote should be recorded this way, since some people may not want their vote recorded this way or may not have a smart phone or social media account. I’m simply asking whether this might work for some — apparently non-trivial, given the popularity of selfies in the voting booth — portion of our population.

I’ve fallen behind in my posts and today I’m going to try to write two. This first one deals with our seminar last week on “AI, the Internet, and the Intelligence Singularity — will the machines need us?” We spent quite a bit of our time together discussing the AI singularity, but I’m going to focus here on the current rapid pace of change in artificial intelligence. It’s fun to imagine what it might be like to interact with a clearly intelligent machine, but as you can see from the students’ blog posts, it is really hard to come to consensus on what each of us would characterize as a clearly intelligent machine. And without consensus, our minds just run wild in talking about what The Singularity — whether a point in time or a process over time — would look like.

With less imagination, what fascinates me is the practical advance of artificial intelligence in our daily lives. I have lived through several cycles of hype around how artificial intelligence would radically change our daily lives, and for the first time, it feels like it is finally happening. Siri was fun when it first came out, but it didn’t change my life and I never really used it. But about a month ago, my family got an Amazon Echo, and Alexa has changed our lives. While it is not perfect, we use it constantly. As a childhood fan of Star Trek (the original series), I feel like I have what Captain Kirk had when talking to the Starship Enterprise’s computer. Wow!

Outside the home, I’m astounded by the rapid adoption of self-driving technology. As someone who still drives a stick shift, I can’t say that I’ll be an early adopter, but I can’t deny that broad adoption of the technology is coming. And coming soon. In the New York Times today, the most emailed article is titled, “Self-Driving Truck’s First Mission: A 120-Mile Beer Run.” Perhaps it’s the reference to beer, but I would bet that this just shows how interested the general public is in self-driving technology. This particular technology comes out of Otto, which is owned by Uber, and founded by researchers from Google’s multi-year efforts into autonomous vehicles. Self-driving trucks is not just a research idea. It’s a business plan for Uber.

And the government has noticed too. About a month ago, the Times wrote an article titled, “Self-Driving Cars Gain Powerful Ally: The Government.” This is an important first step toward a future where our policies, regulations, and laws begin to catch up with the changes that technology is making on the nation’s highways and roads. It will be interesting to watch the battle between oversight and overregulation. And it’s good to see an early push to consider issues of safety, security, and privacy. Too often these issues have been left as an afterthought.

As we talked about in our seminar, self-driving cars are not just a technological challenge. In designing and coding for these cars, software engineers are making ethical decisions. How do you write the code that decides between an outcome that causes a car to swerve and hit a pedestrian and another that causes the car to swerve and injury the passenger? What software engineering practices do you put in place for situations that the artificial intelligence might encounter that are not as obvious to the designer as the example I just mentioned? Working at a college focused on a liberal arts and sciences approach to education, we need the humanities to be as strong as — and interacting with — our engineering programs as we enter this world of ubiquitous artificial intelligence.

This week in class we talked about the Internet of Things (IoT), and I opened up our discussion by asking what kinds of devices the students had in their homes that connected to the Internet. I meant non-computer, non-tablet, non-cell-phone-like devices. Things that you’d look at and think, “That’s a home appliance, not a computing or network device.”

To my surprise, no one spoke up. In fact, I got a lot of blank stares.

Let’s contrast this with our discussion the previous week, where we talked about how the Internet had changed business. How did the students get their music? Off the Internet, even when driving. I, in contrast, still listen to FM radio. Had they used ride-hailing apps? Oh yes. We had reviewed statistics that had shown that users of ride-sharing services are predominantly less than 45 years old. Ok, I’m a bit older than that, and I also still prefer to use public transportation to get around Boston (i.e., buses and the T).

The IoT, it appears, is mostly invisible to our younger generation. Why, I began to wonder, do I want to control my thermostat over the Internet, but I’m perfectly happy to walk 0.8 miles to wait in the T station for the next train to arrive? Why do my kids expect Lyft to show up instantaneously and outside our front door, but probably can’t find the thermostat inside that same house?

At first I thought the answer might be related to our age and place in life. I’ve used public transportation for many years, and putting aside the recent failures of the T to run when we really need it, public transportation has served me just fine. Why do I need to change? But wait, my thermostat has worked perfectly well and very reliably for years too. Why would I want to replace it with one that might have software bugs? This line of reasoning didn’t seem to get me anywhere.

When you’re out of ideas what do you do? Yup, I typed “Who buys Nest products” into Google. I had hoped that this search would provide me with demographic information on those consumers buying Nest products, but other than the top hits directing me to Nest’s homepage, its online store page, and its “Where to buy” page, I was directed to pages discussing Google’s original acquisition of Nest and a couple of pages reviewing the performance of the acquisition two years later. While scanning these articles, it hit me: Nest sells gadgets.

Obvious, I know, but think about it. I grew up in a hardware-dominated generation. Yes, software existed, but the hardware mattered more. Today’s students have grown up in a software world. It’s an app that gets Lyft to show up. Yes, there’s an app for the Nest Thermostat, but you have to buy the thermostat to ever have any interest in getting the app.

It was also interesting to read the contrast between Google buying Nest for $3.2B in 2014 and Apple’s acquisition of Beats for $3B also in 2014. Both are hardware companies, but Beats sells accessories for your Internet-connected phone. Nest sells home appliances that connect to the Internet and simply use the phone for remotely controlling the device. Today, Beats is booming and Nest is viewed as an acquisition failure for Google.

The bottom line here is that I, with my hacked together “Internet-connected home,” am probably not a good example of the immediate future for IoT-focused, consumer companies. The only people I know with these devices in their homes are those who have received birthday gifts from me lately (and employed me as their IT help). The Wired article by Daniel Burrus that we read for this week’s class said, “[w]hen we start making things intelligent, it’s going to be a major engine for creating new products and new services.” A lot of things are getting smarter, but it feels like we’re a ways off from the kind of consumer success we’re seeing in the app world.