Artificial Intelligence’s progress is hard to track, but the news does paint a clear picture. The last few weeks have shown us, once again, how artificial intelligence (AI) is making inroads into sectors and industries that were previously considered far-fetched for this branch of computer science.
From AI’s ability to control a nuclear fusion reactor to AI predicting which company will make the most revenue, the scope of AI is far exceeding its capabilities previously capitulated by data scientists. With more research and more data, the progress of AI is happening faster than before. Here is a look at some of the stories that depict the state of AI right now.
The biggest news came from DeepMind, the Alphabet-owned AI lab, with their ability to unlock another real-world potential for artificial intelligence. Scientists at the London-based AI lab announced last week that they have trained an AI system to control and sculpt a superheated plasma inside a nuclear fusion reactor.
Nuclear fusion is a process that releases vast amounts of energy by smashing and fusing hydrogen, and is touted to offer limitless sources of clean energy. However, the challenges are aplenty and DeepMind’s AI could deliver the breakthrough that scientists are looking for to recreate the nuclear fusion reaction occurring in outer space.
After the astonishing amount of revenue declared by big tech giants during their earnings announcements for the last quarter of 2021, it would be rather difficult to predict which company will make a surprising amount of revenue. Not for AI, though.
Asset manager NN Investment Partners (NNIP) has developed an AI-based smart software that can predict which companies will achieve surprising revenue in the next quarter. The software uses AI’s ability to analyse text to reach a conclusion. NNIP thinks it can increase the return for shareholders with its AI-powered tool.
Another scientific breakthrough for artificial intelligence came in the form of its ability to discover a new pattern in knot theory. Knot theory is a part of mathematics that can be applied in molecular genetics.
In a first for theoretical mathematics, machine learning was successfully used to find mathematical connections in knot theory. This was possible after computers were fed with large data sets with the hope to reveal patterns. The self-learning algorithms show how AI can be used in the field of pure scientific research.
A top researcher at OpenAI, the California-based AI research institute, has claimed that AI may already be gaining consciousness. The startling statement from Ilya Sutskever, chief scientist of the OpenAI research group, shows that AI’s progress is beyond human understanding. Sutskever claimed on Twitter that “today’s large neural networks are slightly conscious”. It is not only an unusual claim but shows that AI researchers could be warming up to the idea of today’s AI not being far from human intelligence.
Renewi Belgium has announced an investment in smart trucks that can separate waste at the source. These trucks are capable of doing waste separation at the source with the help of “smart” cameras. These cameras are capable of identifying different streams of waste using the camera feed and by studying data using AI.
The company also announced that the first 25 Renewi Smart trucks are now driving around. These cameras, beyond identifying different waste streams, help Renewi to comply with the new rules of the eighth edition of the Flemish regulation for the sustainable management of material cycles and waste (VLAREMA).
A team of experts in AI and animal ecology are now proposing a multidisciplinary approach to improve wildlife research. The study published in Nature Communications proposes effective use of the vast amount of data being collected using the new technology.
With the help of satellites, drones and equipment such as automatic cameras and sensors placed on animals or in their environment, there is an unprecedented amount of data studying populations of wild animals. The study proposes development of more accurate models by combining advancements in computer vision with the expertise of ecologists.
This shouldn’t surprise anyone. Smart algorithms are everywhere but these algorithms need to be checked for discrimination and arbitrariness. Dutch Data Protection Authority (AP), the Dutch privacy watchdog, is responsible for keeping a check on these algorithms. However, there is a concern around its ability.
A tour of experts by NU.nl shows that there is a lingering question of whether the Dutch privacy watchdog can handle this task since it is already overloaded. The Dutch Data Protection Authority is constantly calling for an insufficient budget to monitor the privacy laws and from 2023, it will have to check whether algorithms are transparent and ensure that they do not discriminate.