UA RU EN

MIT Research Reveals AI Now Outperforming Human Experts in Complex Evaluations

Дослідження MIT свідчить про те, що штучний інтелект досягає нових висот у складних оцінках, перевершуючи професіоналів.

Artificial Intelligence and Its Impact on Expert Roles

In a recent discussion, AI expert Volodymyr Bandura spoke with political scientist Yuriy Romanenko about the growing capabilities of artificial intelligence. Bandura presented findings indicating that AI models are now achieving superior results on complex tests compared to many seasoned professionals. He referenced data from OpenAI and specialized benchmarks like GAIA/GPQA, which feature intricate practical problems spanning fields from engineering to marketing.

The GAIA/GPQA tests were developed by a team of hundreds of experts, each possessing between 10 to 20 years of industry experience at leading companies. Bandura noted that the latest AI models frequently outperform these highly experienced human specialists in many scenarios. This trend reflects the rapid acceleration of AI capabilities, moving beyond theoretical tasks to practical applications.

'The current top models... they are already scoring a bit better than a good expert in most tasks. Therefore, I completely disagree with the notion that they will fail significantly somewhere.' – Volodymyr Bandura

AI Integration in the Business World

Bandura also commented on an MIT study from 2024 or early 2025 concerning AI implementation in business. He emphasized that the research was conducted by a specialized MIT unit that developed a product based on AI models, describing it as 'beautiful promotional material.'

'This study, it's from '24, or the beginning of '25. So it's already outdated, simply put.' – Volodymyr Bandura

According to the expert, integrating AI tools like ChatGPT boosts the personal productivity of nearly everyone who uses them. Bandura further highlighted a common bias, where many people perceive AI not as an opportunity but as a threat. This perception gap is a significant hurdle for broader adoption, even as the technology proves its utility in enhancing human work.

The advancement of artificial intelligence and its ability to surpass experts in specific tasks underscores the critical need for businesses and society to adapt. Given AI's potential to increase productivity, shifting the public perception from viewing it as a threat to recognizing it as an opportunity is essential. This shift can unlock new horizons for innovation across various sectors, though it requires a thoughtful and informed approach to AI implementation and use.