What keeps Google’s CEO up at night? The dark side of AI advancement: Google CEO Sundar Pichai, expressed his concerns about the dark side of artificial intelligence, if not regulated and deployed appropriately, “It can be very harmful if deployed wrongly and we don’t have all the answers there yet — and the technology is moving fast. So does that keep me up at night? Absolutely,” Pichai said in an interview with CBS’s 60 Minutes.

Should AI be treated like nuclear power? Pichai argued that governments and AI developing companies should consider nuclear arms-style treaties when regulating AI development. The reason for such suggestions stems from the dangers Ai poses to society globally, including disinformation and the technology’s impacts on knowledge-based jobs.

A gap between us and AI and potential safety concerns: There’s also a “mismatch” between users’ thinking and adaptability using AI, and the rapid development pace of AI, Pichai said, calling for a global regulatory framework for AI developers amid fierce competition that could lead to cutbacks on safety measures.

Of course, Google’s tools are safer: The version rolled out for public testing of Google’s chatbot Bard was safe, according to Pichai. Google has been cautious, holding back more advanced versions of the chatbot for public testing, he said.

Remember: Last month, Elon Musk, Steve Wozniak, and Yuval Noah Harari were among 1k who signed an open letter demanding a temporary pause on the advancement of ChatGPT-like programs, warning “human-competitive intelligence” could cause “dramatic economic and political disruptions” to society.

The rise of AI might be worrisome, but it’s tough to deny it’s helpful: Customer service representatives at a Fortune 500 software firm who were given access to generative AI tools were 14% more productive on average than those who were not, with the least skilled employees benefiting the most, according to a recent study by Stanford University and MIT researchers. Researchers tracked over 5k customer support agents at an unnamed Philippine-based software company, which serves small and medium-sized US businesses, measuring the employees’ performance using key metrics like the effectiveness of resolving customer issues successfully in a timely manner.

Least-skilled employees had the most to gain: The study demonstrated that the firm’s least-skilled customer service agents were able to complete their work 35% faster with AI assistance. The improvements in productivity may be due to AI tools’ ability to disseminate the knowledge of top performing workers to less experienced workers via AI-generated suggested responses. This suggests that companies should acknowledge the expertise of their top-performing employees as their knowledge and skills are likely to be used as the foundation for AI tools that will be used across the organizations. This raises the question about how these high-skilled employees should be compensated for their additional service to the firm, MIT researcher and one of the study’s co-authors Lindsey Raymond suggested, according to Bloomberg.

What makes this study different from others? The study is the first to measure the impact of generative AI outside of a laboratory setting. Previous studies have shown the potential of large language models in fields in the workplace, but their true impact has mostly been speculative until tested in real-world settings, Director of the Digital Economy Lab at Stanford Institute for Human-Centered AI and one of the study’s co-authors Erik Brynjolfsson told Bloomberg.

Tags: