AI’s recent rise to prominence has made some wonder if we’re on the verge of an intelligence explosion. But to think that AI just woke up one day and got smarter is a misconception. It’s the culmination of centuries of technological, scientific and philosophical breakthroughs that have led to the intelligent systems we know today.
A new development in AI could bring it closer to human-like thinking, without needing supercomputers or endless data. Scientists have developed a model called Lp-Convolution that helps machines focus on important parts of images — much like the brain does — and can adjust their focus dynamically. This will make AI faster and more effective in real-world applications, such as detecting cancerous cells on scans or X-rays.
In 2024, Google DeepMind created an AI system called AlphaGeometry that was capable of solving complex geometry problems at a level approaching gold medalists in the International Mathematical Olympiad. It was a significant milestone that brought us closer to the dream of AI systems that can discover and verify new knowledge.
An earlier milestone was reached in 2023 when a machine learning algorithm named GraphCast was able to predict weather conditions more accurately than the industry benchmark HRES. This advancement could help companies plan ahead for inclement weather, reducing disruptions and saving money.
While many experts acknowledge AI’s incredible potential, they also recognize that powerful systems have the potential to make humans obsolete and worsen economic inequality. In addition, some are concerned that AI is susceptible to replicating societal biases, and has the ability to discriminate against certain social groups. However, these concerns are being addressed through initiatives to ensure ethical and transparent AI development.