Opinion
Think it over: What AI does for us versus what it’s doing to us
Summary
- Chatting with AI chatbots can leave a neural imprint on us. A friendly bot may make us feel less vulnerable, offering us a sense of comfort, but might an emotional vacuum created by such a relationship also leave us emotionally fragile?
To erase the line between man and machine is to obscure the line between men and gods," says the tagline in the teaser of the 2015 movie Ex-Machina. What technology does for us has long been evident. But, in the last decade or so, ever since social media became central to our lives, what technology does to us has become a matter of profound interest. Even more so now, as conversations on artificial intelligence (AI) go mainstream.
Making machines that behave like or resemble human beings has been a pursuit for about two centuries. As early as the 1800s, Austro-Hungarian inventor Wolfgang von Kempelen tried making a speaking machine using bellows and reeds to model the human voice box. This was later improved by Charles Wheatstone, which in turn influenced Alexander Graham Bell, who tried creating his own device. While Bell’s speaking machine didn’t materialize, it helped him invent the telephone.
The first electronic talking machine, Voder, was created by Bell Labs and exhibited at the 1939 World’s Fair in New York. In the 1950s, early AI pioneers began working on speech recognition (as distinct from understanding). Decades and many breakthroughs later, we are in the throes of a man-machine interaction revolution. But what does it do to us?
Human interactions with ‘intelligent’ machines are not just about machines. They are also about us. When we interact with machines that respond like humans, what does it do to us, our perceptions, emotions and neural wirings?
Also read: Questions have a hot new role in the age of AI: Stirring human minds
Much is unknown, except that it changes us. Just as human-to-human interactions trigger our neurons, make biochemical changes and alter our bodies, so do human interactions with machines. Even as we get into these interactions with the full awareness that they are only machines, we may slowly reach a point where we are ‘in communion’ with them.
The first chatbot came much before the internet went global. An MIT professor and computer scientist, Joseph Weizenbaum, built Eliza in 1966 as a conversational interface between humans and machines. Eliza engaged with users like a psychotherapist and many people started attributing human-like feelings to the programme, wanting to share intimate matters and spending time alone with the machine. AI bots are now being touted as a solution for loneliness, as digital companions. This assumes a chasm crossed from AI’s ability to ‘understand’ what we are saying to an ability to ‘care.’
Is there a Turing Test for empathy? That is the question.
A recent study by Sherry Turkle at MIT, ‘Who Do We Become When We Talk to Machines?’ delves deep into the impact of human–machine conversations. Turkle, a sociologist and trained psychologist, is the author of acclaimed books like The Second Self: Computers and the Human Spirit and Alone Together: Why We Expect More from Technology and Less from Each Other.
She has taught at MIT along with Joseph Weizenbaum and also studied people’s interaction with Eliza, terming our emotional engagement the ‘Eliza Effect.’ Her new MIT study explores how our digital relationships affect our understanding of human connections and is also an attempt to develop methodologies that study how artificial empathy changes our relational capacity.
Turkle coins an interesting term, ‘artificial intimacy.’ The study had participants from a wide range of backgrounds and demographic groups. As either new or regular users of AI-based conversational tools such as Replika, Pi, Woebot and ChatGPT, they were found to experience an emotional bond with the technologies they were engaging with. The result of this study raises some points that might have a far-reaching impact on society and perhaps even on the evolution of our species.
Also read: Getting humans to trust face-less machines remains a challenge for tech giants
First, all or the majority of participants found it easier to deal with machines than humans, because it replaces “stressful human connections." Real relationships have more friction and are relatively difficult to deal with. Talking to machines make people feel less vulnerable, as there is no friction, no second-guessing and no fear of being left behind.
A machine friend (or companion) can offer legitimacy and validation. But wouldn’t a withdrawal from dealing with real people and having real experiences make us more emotionally fragile? As Turkle notes, “If we live in a culture where significant numbers of people say they should never have to be made uncomfortable, and since discomfort, disappointment and challenges are part of most human relationships, that is a dilemma."
Second, consider the argument that AI built on the collective intelligence of many experts is better than any single expert. Eric Schmidt, former chairperson of Google, has been quoted as saying that Generative AI will make “much of human conversations unnecessary." But Turkle asks the most important question: “When we are in human conversation, we often care less about the information an utterance transfers than its tone and emotional intent. In a world that deals in averages, what happens to our sensitivity to all this?"
What does all this mean to us? What does it mean to be human? Is being human now a relative term? Are we being shaped by past data or the law of averages? Will we get more standardized and homogenized?
More questions. Few answers.
