Middle of Somewhere Design

After watching the Star Trek episode “Measure of a Man” and learning about the character Data, it was interesting to compare him to AI powered assistants like ChatGPT, Google Assistant, Apple Siri, and Amazon Alexa. The main difference that I would consider between Data and the other AI assistants is that he is an artificial intelligence that is built as a cyborg-type being that is meant to be mostly indistinguishable from humans in physical appearance. In contrast, the real-life AI assistants are accessed digitally through our electronic devices, like phones, tablets, computes, and smart speakers. Data’s vocal delivery is slightly less human sounding and distinguishable as being different from the other characters. This is supported in the way that his vocal delivery seems somewhat monotone and lacking in emotional inflection. I would say this delivery is what makes him more similar to the real-life AI assistants. In addition, like the real-life assistants, he has processing power that allows him to access infinite amounts of information in almost instantaneous response time.

An interesting aspect of the character Data in Star Trek is that he was expected to be, and seen as, a real crew member of the starship Enterprise. To me, this was the largest difference in expectations when comparing him to real-life AI assistants. I think currently AI like Siri or Alexa are seen as exactly what we’ve been calling them, assistants. They are algorithms and search engines meant to make human life easier by conveying information, completing tasks, and helping solve problems. The character Data can do all these things as well, but the episode frames him as a crew member deserving of trial to defend his own rights and autonomy. The fact that his AI is built into an actual human-appearing body that can travel on missions and be “unplugged” is a major differentiator. The expectations of the real-life AI assistants is to provide information. It can be frustrating when the responses aren’t clear or don’t understand the questions properly. The character Data seemed much more life-like in his interactions when he didn’t understand something, and seemed to learn from mistakes instantaneously, like a real human might.

The uncanny valley theory is an interesting concept that describes a “level of realism in artificial life forms in which the human observer has a negative reaction.” In my opinion, this theory was demonstrated in the Star Trek episode during the trial when Data’s arm was removed and then he was turned off. I think the other crew members had a negative reaction in multiple ways. In one way, it was sort of disturbing because a human-like being (in physical appearance) basically had a limb severed in front of other people. In addition, the act of turning off Data gave the physical appearance of killing him. I don’t think the same type of uncanny valley can occur with a digital AI assistant because it lacks the physical form that conveys realism.

The evolution of both the concept and actual performance of voice interfaces has evolved significantly over time. From 2001: A Space Odyssey in 1968 to Star Trek in 1989, and even Amazon Echo in 2014, there has been large changes in AI voice interfaces. Back when the concept was sci-fi and in our imagination it was connected to robots, then it evolved to human-like cyborgs, and finally digital interfaces in computers. The public’s expectation of voice interfaces has evolved from a movie concept to for fun to today where these assistants are expected to deliver real results and lightning speed and understanding.

As discussed previously, different voice interfaces have different levels of “life” to them. Data differentiates himself from current real-life AI assistants by his human appearance and autonomy. He operates more like a conscious being with free thought and desire to grow. You could make the argument that the AI assistants have free thought and a desire to grow as well, however, to me, the differentiating component is that their algorithms require human interaction and prompts in order to perform their capabilities. It is hard to place a ranking on their personalities or character, but as mentioned the physical body of the character Data puts him at a level that real-life AI assistants of today have not achieved.

The thought of assigning emotions to a voice interface is an interesting debate. Many AI assistants are built to be less emotionally appealing due to the mundane tasks that they are created for. Most of the utilitarian performance would feel unnecessary to incorporate an emotionally appealing voice interface. I guess a situation where it would be more appealing would be creating a voice interface meant for a companion of sorts. The movie Her comes to mind, where Joaquin Phoenix has a voice interface companion that he ends up falling in love with.

Through watching the Star Trek episode and reading more about voice interfaces I think one of the main takeaways is the importance of finding the right balance of utility and emotion necessary for the intended operations and goals of the design. Humans can have strong opinions, as well as expectations, about AI now that it is mainstream. As they currently exist, AI interfaces are thought to be digital assistants, not autonomous beings. I personally find many of them frustrating to use still, and can easily cast them aside, which is very different than how the Enterprise crew felt about losing their comrade, Data.