Articles > AI for Business - Your 5 reads of the week #11

a month ago


Welcome back to your favourite place for AI for Business news. This is our 11th edition of AI Business reads, we are not giving up any time soon, so keep coming back each week! We hope you continue to enjoy these reads and let’s dive in the 5 hand-picked articles we have for you this week!

This week you will read about:

  • Google's lung cancer AI outperforms experts
  • US Postal invests in self-driving trucks
  • World's first raspberry-picking robot set to work
  • Textract, Amazon's machine learning parsing application
  • The future of AI is collaborative


57% of companies expect AI to have a high impact or a very high impact on business areas that are “entirely unknown to the company today”. The machine power behind AI gives companies the opportunity to analyze more data and thus get more insights about their operations, customers and others. This might enable them to invest in new business areas in order to respond on demands of customers they did not know existed before thorough analysis of internal data.


According to the World Health Organization, lung cancer is one of the most common causes of death globally with 2 million people dying annually. Google AI researchers made a deep learning model that can detect lung cancer 5% more often on average than a group of 6 human experts. The model is trained with 42 000 CT scans from 15 000 different patients. It is now able to predict whether a patient has lung cancer, generate a risk score of bad tissue and identify the location of this bad tissue that is causing lung cancer.

The postal industry has to cope with two structural shifts within their industry. On the one hand, the traditional core business of letter delivery is in decline as communications move online. On the other, the industry faces fierce competition in the rapidly-growing e-commerce parcel market. Hence the traditional postal companies need to continue to innovate. This week the US Postal Office started to use self-driving trucks between Phoenix and Dallas, this to improve service, reduce emissions and save money. Self-driving trucks are the logical consequence of the national shortage of truck drivers. For now, the trucks will have a safety driver and engineer on board in order to monitor the autonomous systems.

In the UK, the farming industry battles rising labour cost and a shortage of seasonal workers due to the Brexit. In order to battle these shortages, farmers are investing in automation, one example in the berry robot from Hall Hunter (one of Britain's main berry growers). The robot uses sensors and 3D cameras to recognize the berry and thanks to machine learning is able to pick the berry, this entire process takes 1 minute. But since robots do not have to take breaks they outperform the 8-hour shifts from humans by 10 000 berries a day. The biggest challenge for the robot is the light since this makes it more difficult for the 3D cameras to recognize the berries.

The Department of Computer Science at the University of Copenhagen has built an AI model to detect whether a student has written an assignment on their own or had it composed by a ghostwriter. Due to the rise of online platforms providing writing services, schools are looking for new tools to combat this new type of cheating. Schools are already using anti-plagiarism platforms to find out if certain parts are copied directly from a previously submitted paper, but ghostwriters are more difficult to intercept.

The model is based on analyses of 130 000 written Danish assignments and can predict with 90% certainty if an assignment was composed by a ghostwriter. This by comparing this work with previous works and looking at the length, structure of sentences and the use of certain words.


Artificial Intelligence is too often linked with replacing jobs, the people making this link often tend to ignore the collaborative and job-creating attributes of AI. Researchers from the universities of Göttingen, Germany, have found that a team that consists of machines and humans outperform those with just humans or just machines. AI should be seen as a tool that analyzes the huge amounts of data a company gather and find opportunities/insights in it that humans might have overlooked. And since these AI models are being more and more complex, the humans are needed to explain these findings and act appropriately.

That's it for our 11th edition! We hope you enjoyed the readings! If you missed some of our earlier must-read lists, head over to our Articles section to catch up!

And do not forget to subscribe to our newsletter for your monthly AI portion.

See you next week!