A new generation of AI assistants, powered by advanced models such as GPT-4, is attracting billions in investment from Silicon Valley. These assistants, often referred to as “agents” or “copilots,” are expected to perform more complex personal and professional tasks with less human supervision.
Div Garg, a developer at MultiOn, envisions this new breed of AI to evolve into something akin to Jarvis, Tony Stark’s AI assistant from the Iron Man franchise. However, according to Kanjun Qiu, CEO of Generally Intelligent, complex tasks that require reasoning skills, such as scheduling meetings, are still challenging for AI, Reuters reported
Tech companies and startups are investing in the development of these increasingly autonomous AI agents, aiming to create artificial general intelligence (AGI) capable of matching or surpassing human cognitive abilities. Inflection AI, which raised $1.3 billion in late June, is working on a personal assistant capable of handling tasks like securing flight credits and hotel bookings in case of travel delays.
The race towards more autonomous AI agents was accelerated by the release of GPT-4 by OpenAI. This advanced model facilitates strategic and adaptable thinking, vital for navigating real-world unpredictability.
While early versions of these agents show promise, developers warn that they can also make mistakes, like wiping a computer’s drive or making incorrect purchases. Aravind Srinivas, CEO of Perplexity AI, likens AI to a baby that needs constant supervision.
The emergence of increasingly independent AI agents has also raised ethical concerns. Yoshua Bengio, a leading figure in AI research, earlier warns that future versions of the technology could act unpredictably and potentially harm people.
The commercial potential for AI agents is large, attracting interest from major tech companies like Microsoft and Google. Microsoft CEO Satya Nadella sees foundation-model technology as a significant advancement from existing digital assistants like Cortana, Alexa, Siri, and Google Assistant.
However, the first iterations of quasi-autonomous AI agents might come from agile startups rather than established tech giants. A surge of investor interest in this field is fueling the commercialization of AI agents, with at least 100 serious projects currently underway.
While the potential for AI agents is exciting, developers and investors understand they are more than just advanced chatbots. As this new wave of AI continues to evolve, it seems we are heading towards a future where AI assistants are capable of more complex and autonomous decision-making.