As companies prioritize accuracy, AI discussions at Davos center on generative technology.

• Discussions about generative AI have dominated both private and public talks at the World Economic Forum in Davos, Switzerland.
• Intel CEO Pat Gelsinger stated in an interview that higher generative AI accuracy is necessary to advance technology into new applications.
• Salesforce's CEO of AI stated in an interview that the use of human co-pilots and confidence level testing would enhance the reliability of the technology.

The dominance of generative artificial intelligence in discussions at the World Economic Forum has been highlighted by the temporary takeover of local storefronts by major technology companies, including Google and Microsoft, to demonstrate their power to conference attendees.

In a CNBC interview, Chief Executive Officer Pat Gelsinger stated that while 2023 showcased the technology's capabilities, 2024 will focus on enhancing the accuracy of results to enable executives in high-stakes industries such as hospitals and manufacturing to feel confident in utilizing AI.

Gelsinger stated, "The conclusion of today's AI utility has been reached. In my opinion, the next phase of AI will focus on incorporating formal correctness into the foundational models."

Gelsinger stated that humans must become more accustomed to the reliability of technology, whether it's doctors utilizing artificial general intelligence (AGI) for diagnoses, warehouses employing it to detect assembly line issues, or self-driving vehicles.

Gelsinger stated that while AI has successfully solved certain problems such as basic prediction, detection, and visual language, there are still many unsolved issues. Proving the accuracy of large language models is one of these challenges, as there are numerous errors present today. Despite this, Gelsinger emphasized the importance of improving the productivity of knowledge workers. However, he acknowledged that the ultimate goal is to have the knowledge worker verify the accuracy of the AI solution.

To enhance accuracy, it is recommended to conduct experimentation and co-piloting tests to promote adoption of AI, according to Clara Shih, CEO of Salesforce AI. The AI technology can adapt to varying levels of confidence as users become more comfortable trusting it in high-stakes situations, Shih stated in an interview.

Shih stated that three phases will guide the adoption of AGI. In the first phase, the technology will be used as an assist for work. In the second phase, the technology will be monitored in its autopilot mode to ensure accuracy. Finally, the technology will be trusted to work at the desired confidence level, Shih said.

Shih explained that the AI can be instructed to be more cautious in higher-risk situations until it is essentially "graduated" to autopilot mode by a human co-pilot.

Open AI CEO Sam Altman stated during a panel discussion with Bloomberg's Brad Stone that the three-phase approach, which heavily relies on human acceptance of the technology's reliability, should make AGI less frightening than some have predicted.

This tool is much more than I anticipated, and while it will improve, it's not yet replacing jobs. It's an incredible tool for productivity that enhances human capabilities and allows AI to handle certain tasks.

In an interview, CEO Adena Friedman stated that the past year has been "a year of discovery." The financial industry, including Nasdaq, will employ AI to modernize outdated systems, streamline automated workflows, and save employees time each day, according to Friedman.

A year ago, it emerged on the scene," Friedman stated. "We conducted some experiments and recognized its potential. This year will be the year of activations for us and everyone.