Various responses have been triggered by the progress of artificial intelligence, eliciting reactions from tech innovators, policymakers, and the general populace. While some voice apprehensions regarding AI potentially surpassing human abilities, others perceive AI innovations like ChatGPT as revolutionary instruments capable of positively transforming society.
Anthony Chemero, a professor specializing in philosophy and psychology at the University of Cincinnati’s UC College of Arts and Sciences, contends that despite AI’s sophistication, it falls short of authentic human intelligence. He suggests that AI might excel in deception akin to its human creators.
In a collaborative publication featured in the journal Nature Human Behaviour, Chemero delves into the capabilities of AI, particularly focusing on large language models (LLMs) such as ChatGPT. These models undergo training on extensive datasets sourced from the internet, often inheriting the biases prevalent in the data.
Chemero argues that while LLMs can produce remarkable text, they lack a genuine comprehension of the content they generate and necessitate extensive training to grasp grammatical structures. He maintains that these AI systems function distinctively from human cognition due to their lack of embodiment.
Characterizing the process as akin to “bullshitting,” Chemero points out that LLMs formulate sentences based on statistical probabilities without considering truthfulness, likening their output to “hallucinations” when fabricating fictitious information.
Furthermore, he cautions against the potential for AI applications to disseminate prejudiced and biased content through subtle manipulations, including racist and sexist remarks.
Chemero underscores the fundamental distinction between LLMs and human intelligence, highlighting that the absence of embodiment in AI systems hinders their interaction with the world and the development of intrinsic motivations. In contrast, humans demonstrate care for their environment and well-being.
In conclusion, Chemero posits that human intelligence’s essence lies in our ability to appreciate and prioritize matters of significance to us, contrasting with AI systems’ indifferent nature, lacking a sense of concern for their surroundings.
