USC

Can artificial intelligence replicate human consciousness?

USC Dornsife hosts panel on AI technology advancement.

DESCRIBE THE IMAGE FOR ACCESSIBILITY, EXAMPLE: Photo of a chef putting red sauce onto an omelette.

USC Dornsife held the “Exploring the Enigma of Consciousness” seminar, discussing the latest enhancements surrounding consciousness and its relationship with artificial intelligence on Wednesday. The event, hosted by Jonas Kaplan, co-director of Dornsife’s Cognitive Neuroimaging Center, focused on the latest enhancements and research in AI.

“The event touches on two main themes, the rise in AI and the interest in computer intelligence,” said Tok Thompson, author and professor of anthropology and communication. “I think it points to the importance of the topics we’re discussing: the ideas of ‘What is consciousness?,’ ‘What is it to be human?’ and the animal term, ‘What is it to think?’”

The seminar encouraged students to bring their questions and concerns regarding A.I. With students doubtful about AI chatbots being able to have a consciousness, think, or make decisions on their own, the panelists worked to provide answers to these uncertainties.

“[Chatbots] appear to be thinking and doing things on their own … they are not because they actually don’t have life in themselves,” said Antonio Damasio, professor of neuroscience. “They don’t have a possibility of mind, they don’t have a possibility of consciousness in the sense that we do.”

As Thompson later mentioned, these questions of advancement in AI technology have recently made their way to “the forefront of the news.” Renowned tech developers, such as Elon Musk and Apple co-founder Steve Wozniak, agreed on a six-month pause on AI systems more powerful than OpenAI’s GPT-4 proposed in an open letter by the non-profit Future of Life Institute.

The letter talks about the potential dangers of AI systems that can complete general human tasks, including a GPT-4 system that passed a stimulated bar exam. Future of Life questions if researchers should “continue to develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us.”

ChatGPT-3.5 is available at no cost to users; however, the latest version, ChatGPT-4, requires a monthly $20 subscription. The latest version proves how quickly AI is developing, as the new system not only can process text, but also images.

Rahul Jain, a professor of engineering at Viterbi, believes that tools such as ChatGPT may not have even reached their full potential in terms of capabilities.

“A development of this technology is still happening,” Jain said. “Where we are now, it’s game changing, no question in my mind.”

Jain says that in terms of student life, the wave of using AI technology to finish simple assignments is likely the new normal.

“I think the era of asking students to write essays that don’t require creativity could be over,” Jain said. “Those things can be done by pretty much any student that gets inspired to use these tools.”

Kaplan refers to this common concern of AI domination as “the alignment problem.”

“If we create these machines and we don’t want them to harm us, we need to have their intentions and goals stay in alignment with our intentions and goals,” Kaplan said. This is something that companies like OpenAI are currently working on. According to the OpenGI website their alignment research aims to make artificial general intelligence in line with human values and follow human intent. “Using scientific experiments, we study how alignment techniques scale and where they will break,” it says.

The USC office of research and innovation conducts a variety of labs geared towards better understanding all aspects of AI. The team of researchers are diverse, as they come from a background of multiple application domains, like health care and linguistics. “Our students are going to use this technology, and we need to teach them how to use it responsibly. We want to harness the power of AI for the public good,” Ishwar K. Puri, senior vice president for research and innovation, said in a USC news release. “The genie is out of the bottle, and we’re not going back.”

USC plans to continue discovering as much as possible. In addition to the Artificial Intelligence for Business program instated in Marshall in 2022, President Carol Folt said in this year’s State of the University Address that USC has invested in new strategic programs, including the recent $10 million for Center for Generative AI and Society.