Cybersecurity analyst Jean Linis-Dinco cautions the public against over-humanizing artificial intelligence in the fourth and last installment of the Philippines Communication Society webinar series, last June 28, 2023.
A cybersecurity analyst has cautioned against the tendency of over-humanizing artificial intelligence (AI) as it can distort people’s understanding of what it can do and lose control over AI technologies.
Jean Linis- Dinco, a University of New South Wales cybersecurity student, made the statement during the international hybrid symposium of the Philippines Communication Society (PSC) on June 28, 2023.
Linis-Dinco said AI only produces texts based on patterns without understanding what it is saying, and cannot think abstractly like what humans do.
“Emily Bender came up with a great metaphor for AI. She called it ‘stochastic parrot.’ So it is just like a parrot that can repeat you in speech without understanding it. They just predict the next likely word based on their training data,” she said in the fourth and last installment of the PSC webinar Series titled “AI, naku! Understanding the Impact of Artificial Intelligence on Media and Communication.”
Linis-Dinco said over-humanizing artificial intelligence could lead down the “wrong path,” affecting policies, how the public sees this technology, and even ethical considerations.
“By misunderstanding AI’s functionality and limits, we risk losing sight of the larger implications of this technology that is happening today,” said Linis-Dinco, who was named one of the “Top 100 Brilliant Women in AI Ethics.”
Linis-Dinco said a comprehensive understanding and critique of AI technologies must take broader issues into account.
Broadcast journalist Kara David moderated the international symposium. Gian Rowelle Erenia