The rise of Artificial Intelligence could be either the best, or the worst thing ever to happen to humanity.” It was back in 2016, when world-renowned physicist and thinker Stephen Hawking again warned against the perils of what he considered an existential threat to mankind, however painting a little less gloomy picture than usual. Although no new kid on the block, in 2019, there seems to be no hotter topic than Artificial Intelligence with researchers and tech companies pushing the boundaries of what is capable with it further and further.
We have seen technologies soaring in hype that delivered on their promises whereas others only provided false hopes and crashed. So, what is all the buzz around AI about? Is it just hype and hope? Or even false hopes?
Our lives are already touched far more by this intelligence than most might believe. We talk to voice assistants and chatbots, Facebook recognizes the faces in the photos we upload. AI plays a significant role in health care. The MIT developed a prediction model which can envisage the development of breast cancer up to five years in advance. A new machine learning algorithm detects signs of Alzheimer’s even up to 9 years before the symptoms are full blown. This is only the tip of the iceberg. In 2025, such techniques will be able to make reliable diagnoses and provide evidence-based therapies. Interaction systems combining sensors, machine learning and man-to-machine-interface techniques will open new possibilities for people with cerebral movement disorders.
“Machine learning, together with robotics and drones and the Internet of Things, will improve the monitoring, understanding and prevention of damage on land, air and water”
With machine learning and deep learning already being used in various areas such as customer service, trading and healthcare, we are only at the dawn of the age of super intelligence. AI has the chance to be one of the most disruptive technologies in history, advancing us not only as individuals but as a society. Imagine that: What if AI had the potential to make a difference between life and death for thousands of people? What if it was able to predict the occurrence of natural disasters or conflicts way in advance before they occur? What if AI contributed massively to diminishing greenhouse gases?
Of course, AI is no silver bullet. However, I truly believe this technology has huge potential to help solving some of the major global challenges we are facing - whether political instabilities, food shortage or climate change - and to put our world on a more sustainable footing.
Machine learning, together with robotics and drones and the Internet of Things, will improve the monitoring, understanding and prevention of damage on land, air and water. Numerous projects are already underway. Google is building a platform that will be able to predict floods and warn people through Google Maps way in advance. IBM’s Watson will be able to predict volcanic eruptions with the help of seismic sensors and geological data. With the usage of such tools, it will also be possible to gain a greater understanding of air pollution which can lead to a more effective implementation of preventive measures. This is a major step forward considering that according to the WHO, more than 90 percent of the children in the world breathe toxic air every day.
AI is also poised to have a huge impact on the CO2 emissions of the transport sector, of which two-thirds is generated by road users. AI is the enabler of autonomous driving and will not only help to reduce the number of accidents by more than a quarter, but also to reduce emission. With autonomous driving, there will be less car usage through the usage of autonomous ride shares. We will spend less time looking for a parking spot, which today causes 30 percent of urban traffic. Rough estimates say that the emission reduction through autonomous driving can amount to 94 percent per mile in the United States by 2030.
The rapid advancements are only possible thanks to more sophisticated algorithms, more computational power and the availability of more data. On the other hand, the data volume produced through AI will increase tremendously. The autonomous car of the future will not run on gasoline, but on data. It is estimated that driverless cars will generate around 40 TB of data every 8 hours of driving. AI applications altogether are roughly producing already 80 exabyte’s per year, by 2025 this number is expected to rise to 845 exabyte’s.
The key piece in that puzzle is electronics, lots of hardware is needed to bring these huge amounts of data to life. Sensors collect the data, the data then needs to be processed quickly by powerful processors, stored securely and transferred without any latency to other elements in the chain. Large computing architectures in even larger data centers are needed to handle the training of neural network models which are the heart of almost any AI application.
To handle the data flood and to deliver the next breakthrough, special AI chips are being developed that can provide parallel processing. Through enhanced speed and performance, they will accelerate the usage of deep learning, natural language processing and other AI applications. The next generation of AI chips will improve the computing performance by 1000 times. The specialty chemicals and technical solutions provided by Merck KGaA, Darmstadt, Germany are an important driver for these advances.
Electronics is unlocking the power and potential of data, opening new possibilities to transform our lives and the planet we are living on, making it a more sustainable place.