“Human intelligence is so highly specialized that it is inappropriate to call it ‘general purpose’. We need extremely diverse assistants.”

This is what Yann LeCun, Meta's chief AI scientist and New York University professor, said at the recent NVIDIA GTC 2025, sharing his vision for the future of AI. He predicted that artificial general intelligence (AGI) would be commercialized within 3 to 5 years and emphasized the importance of developing various AI assistants and open source projects.

Customized advanced artificial intelligence, also known as AI agents or agentic AI, is accelerating its establishment as a digital companion for work and daily life.

At GTC, NVIDIA recently announced that businesses and organizations around the world are deploying AI agents built with NVIDIA AI Enterprise to augment human capabilities and accelerate economic growth.

Whether for humanitarian aid, customer service, or healthcare, AI agents are forming a new digital workforce, helping professionals achieve their goals faster and with greater impact at lower costs.

■ Educational Customized AI Agent Case
AI can instantly translate, summarize, and process multimodal content in hundreds of languages. When integrated into agentic systems, this technology will enable international organizations to communicate and educate stakeholders around the world more effectively.

The United Nations (UN) is working with Accenture to develop a multilingual research agent that supports more than 150 languages to promote regional economic sustainability. The agent acts as a researcher and answers questions related to the UN’s Sustainable Development Goals.

Global nonprofit Mercy Corps has partnered with Cloudera to deploy an AI-based method matcher tool. This tool provides humanitarian professionals in over 40 countries with research, summaries, best practice guidance, and data-driven crisis response to help them deliver aid more quickly in disaster situations.

■ Conquering the CS field with AI agents

AI agents provide 24/7 support, handling common inquiries with more personalized responses, freeing up human agents to solve more complex problems. Companies in a variety of industries, including telecommunications, finance, healthcare, and sports, are benefiting from AI agents.

Intelligent routing capabilities categorize and prioritize customer requests to ensure customers are connected to the right experts, while AI agents’ predictive analytics capabilities anticipate issues and enable agents to provide proactive support with data-driven insights.

AT&T worked with Quantiphi to develop and deploy a new Ask AT&T AI agent in its call centers, reducing call center analytics costs by 84%. Southern California Edison is working with WWT to use the agent to improve data processing and prediction across its 100,000+ network assets.

The bank said that by implementing ServiceNow Dispute Management, built with Visa, it was able to reduce call center workload by up to 28% and shorten the time to resolve disputes by about 30% using AI agents.

The Ottawa Hospital has partnered with Deloitte to deploy a team of patient care agents available 24/7. This will enable us to provide pre-operative support and answer patient questions about their upcoming surgery for over 1.2 million patients across Eastern Ontario, Canada.

■ SAP, “AI Agent, Evolution into a Digital Companion”
“AI agents are now evolving beyond mere tools into digital companions. If AI agents are used well, corporate productivity can be increased by 30-40%.”

Christian Klein, CEO of SAP, said at a press conference held in Korea on the 20th that AI agent innovation is taking place as companies focus on accelerating digital transformation. AI agents analyze data in real time and provide decision support and automation of work processes, thereby improving work productivity.

SAP will release the Joule AI agent in Korea in March of this year, distributing AI agent solutions across the business and providing workflow automation, dispute detection, case verification, and solution suggestions in companies' sales, HR, IT, and finance.

■ AI Agents, New Tools for Security Threats

Global research firm Gartner recently warned of the severity of account takeovers (ATOs) and social engineering attacks due to the advancement of AI technology, and emphasized the importance of preparing security countermeasures. According to Gartner, AI agents will reduce the time it takes to take over accounts by 50% by 2027.

“Hackers who obtain passwords through data breaches, phishing, malware or other means exploit the fact that users reuse the same passwords across multiple sites,” said Jeremy D'Hoinne, VP analyst at Gartner.

AI agents can streamline the entire process of stealing user credentials by leveraging automated bots and deepfake technologies, increasing the risk of account hijacking.

Accordingly, it is necessary to encourage the transition from passwords to multi-device passkeys to strengthen user account security, and security leaders should closely analyze the threats posed by AI technology and prepare response strategies, he said.