The Dawn of a New Scientific Era, Powered by AI
The rapid advancements in artificial intelligence (AI) are poised to reshape scientific discovery, offering the potential to accelerate research and solve complex problems at an unprecedented pace. The latest large language model (LLM) releases, such as Anthropic’s Claude 3.7 and xAI’s Grok 3, are showcasing capabilities that rival those of advanced researchers.
These breakthroughs are pushing toward the vision of a world where everyone has access to a “great polymath,” an AI capable of drawing upon vast information to solve complex problems across disciplines, as former Google CEO Eric Schmidt envisions.
Professor Ethan Mollick of the Wharton Business School highlighted the significance of increased computational power in training these advanced models, with Grok 3 utilizing up to ten times more computing power than GPT-4 at its launch. The result is a substantial leap in capability for these “gen 3” AI models.
Claude 3.7 demonstrates emergent abilities, such as anticipating user needs and considering novel problem-solving angles. Anthropic claims that its model is the first hybrid reasoning model, combining a traditional LLM with advanced reasoning capabilities.
The convergence of increased compute power and advanced problem-solving skills is supercharging AI’s abilities, according to Professor Mollick.
OpenAI has entered the AI research space with the launch of its “deep research” AI agent, which is an impressive tool that can accelerate research, analysis, and other forms of knowledge work, making use of an unreleased reasoning model known as o3. This tool can engage in extended reasoning through a “chain-of-thought” approach, breaking down complicated tasks into multiple logical steps, similar to a human researcher.
Deep research can also search the web, providing access to the most up-to-date information. One test demonstrated the tool’s impressive capabilities by generating a 4,000-word report on building a hydrogen electrolysis plant in just four minutes. As a mechanical engineer estimated, it would have involved an experienced professional a week to create something similar.
Google DeepMind’s “AI co-scientist,” also built on the Gemini 2.0 LLM, aids scientists in creating new hypotheses and research plans. Imperial College London has already proven the value of this tool, with AI replicating findings concerning antibiotic resistance in just 48 hours.
While human scientists are still needed to confirm the AI’s findings, the AI application has the potential to “supercharge science.”
The Future of Scientific Progress
Anthropic CEO Dario Amodei anticipates that “powerful AI” — often referred to as artificial general intelligence (AGI)— will accelerate biological research progress by decades in a much shorter time frame. With the recent advancements in AI models, this vision seems increasingly plausible.
However, biology, which is bound by laboratory validation, regulatory approval, and clinical trials, will still experience constraints in the real world.
OpenAI CEO Sam Altman claims that “systems that start to point to AGI are coming into view”. Altman describes AGI as “a system that can tackle increasingly complex problems, at human level, in many fields.” Altman believes that achieving AGI could unlock a new utopian future, where diseases are cured and humanity is able to realize creative potential.
Navigating the Challenges of an AI-Powered World
AI’s advancements present significant promise, yet the rise of AI hasn’t been without its challenges. Setbacks like the downfall of the Humane AI Pin illustrate the difficultly of real-world AI application, which often face significant obstacles, from technical issues to infrastructure limitations.
The potential of AI, however, should be noted, as scientists can use AI to compress research timelines and unlock opportunities across disparate fields. Scientists have already reduced years of work to months or even days.
Altman himself acknowledges the potential for disruption, warning that “the balance of power between capital and labor could easily get messed up.” This concern is already materializing, with reports of job cuts in some cities along with increased AI investments.
Societies must adapt to the ongoing technological revolution by investing in the ethical development and governance of AI, education, and workforce adaptation to ensure that AI’s benefits are shared equitably. The goal is to allow people and AI to combine efforts and usher in a new age that brings about breakthroughs that were once considered impossible.
