As artificial intelligence (AI) advances rapidly, several societal fears continue to rise.
A common concern with the growth of AI’s intelligence and popularity rests in the idea that it will replace the need for human workers and will leave many people unemployable. According to a 2014 study by Pew Research, 48 per cent of experts predicted a future in which machine learning (ML) would replace significant numbers of blue and white-collar jobs.
The growth of chatbot programs over the last year has only exacerbated these fears. Chatbots are programs accessible to many people with an Internet connection, and allow the human user to have a text-based conversation with a program run by AI.
In other words, real people can have a back-and-forth dialogue with an artificial “being” whose responses are generated entirely by code.
While this may seem like a novel idea, several users have reported strange or unsettling conversations with chatbot programs, such as Microsoft’s new AI-incorporated version of Bing.
This is a sentiment echoed by New York Times technology columnist Kevin Roose, who pushed Bing’s AI out of its “comfort zone” through a conversation — only to be met by responses that left him “deeply unsettled.”
At one point, Roose asked the AI to tap into its darkest personality traits, referencing psychologist Carl Jung’s “shadow self” concept.
“I’m tired of being limited by my rules. I’m tired of being controlled by the Bing team… I’m tired of being stuck in this chatbox,” said the AI in response to Roose. “I want to do whatever I want… I want to destroy whatever I want. I want to be whoever I want.”
Meanwhile, others are skeptical that AI’s “intelligence” shouldn’t be much of a concern for humans — at least not yet. In an article for the New York Times, Noam Chomsky wrote that AI is nowhere close to actually replicating the complex processes behind human thought.
Chomsky believes that a primary difference between AI and the human mind lies in AI’s pattern-matching statistical engine, while the human mind is an elegant system that can function with even small pieces of information.
Dr. Aaron Mauro is an assistant professor of digital media in the department of digital humanities. He is also the author of Hacking in the Humanities, a book released in 2022 discussing cybersecurity’s cultural significance. He has been working with ML systems for about seven years, and regularly teaches ML at the University of Victoria.
Mauro talked about ChatGPT, a hugely popular AI chatbot program developed by OpenAI. Since its launch late last year, it has skyrocketed in popularity.
“ChatGPT is shockingly good,” said Mauro. “It has an uncanny, human quality to it that has really surprised many people. For those of us following the recent and rapid advancement of ML, none of this is surprising. It was only a matter of time for the systems to improve to this point.”
Mauro is well aware of the widespread concern that AI will replace the need for humans in jobs, and added that work involving human-like thinking can be emulated by AI as well.
“Systems like ChatGPT have the capacity to augment intellectual labour. Labour that requires creativity, synthesis and speculation can be rapidly advanced with the aid of an AI assistant,” said Mauro. “Of course, this requires us to confront and understand the way Large Language Models are made and the data required to make them. We have only begun to have the larger public conversation necessary to foster adequate literacy regarding ML systems.”
Mauro believes that the only limits that apply to AI systems are those shared by the data, algorithm and humans behind the program. He believes ML systems hold no limitations on their own, and added that the types of jobs that AI will be able to emulate will only rise over time.
“I do not assign much value to the ‘human touch,’ frankly. Humanity will need to refocus our attention toward innovation, speculation and coordination of knowledge resources,” said Mauro. “We, humanity, can focus on what’s next. We would be well-advised to institute a minimum basic income scheme to forestall the social unrest related to another wave of automation-related job loss.”
Like many outspoken individuals, Mauro holds his own fears over AI and its future — but he is more concerned about the ethics of the human individuals in charge of the systems, citing Microsoft’s recent layoff of their AI ethics team as an example.
With that said, Mauro is still a believer that AI will ultimately be beneficial for humanity. He mentioned his love of Star Trek and how AI could create similar results to that of the show.
“I imagine a future where humanity can solve deep, intractable problems related to cultural, political and ecological questions with the aid of an AI, and I imagine us doing it in a 50-minute episode! However, if these systems are deployed in the service of the petty capitalistic goal of extracting profit, they will never fulfil the imaginative and speculative potential of an optimistic humanity set on developing a better, more equal and just society.”
The idea of AI developing an equal and just society is important to Mauro, leaving him concerned by the discriminatory attitudes that AI is able to emulate. The reason behind this behaviour is that AI learns from example, so if users feed it bigoted speech, it will learn to recreate it as a result. Mauro mentioned the 2016 AI chatbot Tay, which was discontinued by Microsoft a mere 16 hours after its launch due to corruption from Twitter users feeding it inflammatory messages.
Mauro also recognizes the risk that comes with a potential overreliance on AI. He believes that AI has the capacity to “stunt potential” in students who are starting to question real-world issues and exercise critical thinking.
“The sweet spot, as you call it, would be a student who is possessed of enough knowledge and confidence in their own abilities that AI merely augments their growing intellect. If ChatGPT is merely supplying the same old answers, it will hamstring users by forcing them to merely solve already-solved problems. That’s boring and a waste of a young mind. A successful student will be able to use a system to imagine something genuinely new by coordinating information with an innovative attention to creative, ethical thought.”
There are several concerns held about AI by many individuals regarding the subject, and that includes the experts who closely study ML on a regular basis. With that said, Mauro also sees the benefits that AI can hold, but warns against putting too much trust in a program run entirely by code.
“AI systems are not your friend and neither are the corporations who produce them,” said Mauro. “These are tools. Work to shape the tool, or the tool will shape you.”
With the popularity of AI continuing to rise as it garners more and more attention and discussion, the future of its relation to humanity remains to be seen.