Case Study

Artificial Intelligence alongside journalists

The new collaborative writing tool developed with Machine Learning Reply wins the SMAU Innovation Award.

#Artificial Intelligence
#Machine Learning

The synergy between creativity and Artificial Intelligence

Machine Learning Reply, in collaboration with the Intesa Sanpaolo Innovation Centre team (through its Artificial Intelligence Lab), and the ISI Foundation, has developed an innovative Artificial Intelligence tool for Portobello, an Italian media bartering company.

The tool, which won the 2021 SMAU Innovation Award represents an important example of cooperation between the creativity of writing and AI’s ability to collect and process thousands of pieces of information in a very short period of time.

The initial objective

Portobello is an Italian media bartering company that uses advertising spaces to buy products from brands, before reselling them in Portobello stores at highly discounted prices. The company’s business model therefore requires a large number of advertising spaces, as well as related quality content in magazines and websites in which to insert these spaces. In order to increase its production capacity in terms of content generation, Portobello has launched an innovative project that involves the creation of an AI tool that helps journalists with writing articles.

The functioning of the tool

The tool provides for a multi-stage interaction with the journalist, who is not only able to select the specific topic, but also work together with the software to write articles, impacting both the writing style and the choice of specific content. The pipeline for generating an article consists of the following phases:

  • the journalist defines web search-style keywords for the article and the software performs a global search of all the websites and newspapers that have covered that specific piece of news, classifying them in terms of reliability and pertinence;

  • the software analyses the text, then proceeds to catalogue the recurring sub-topics in the articles and clusters them, thus allowing the journalist to choose which “clusters” – or thematic groups – to include in the generated text; 

  • the final phase consists of the actual reworking of the text and in writing a completely new article, with the journalist having final control over the text.

The tool also includes a control step, which is made possible by algorithms catalogued in AI as “explainability”. These algorithms provide for the implementation of techniques that allow the generated content to be traced, sentence by sentence, to the corresponding sentences in the sources, thus enabling a prompt verification of the content and therefore its accuracy.

The tool enables journalists to improve their productivity, by allowing them to spend more time investigating stories. Portobello can thus increase the amount of content produced and the consequent sale of related advertising space. In the future, the system can also be extended to include the creation of content for companies’ social media pages.

Picture

Portobello is an Italian company, founded in Rome in 2016. It is a leader in media bartering, namely the use of advertising space in exchange for products of the brands advertised, which are then sold at discounted prices. Portobello’s media offer includes outdoor spaces located in areas of high traffic, video walls, magazines owned by the company and much more.

Picture

Machine Learning Reply is the Reply Group company that specialises in Machine Learning, Cognitive Computing and Artificial Intelligence solutions based on the Google technology stack. Based on the latest developments in artificial intelligence, Machine Learning Reply applies innovative Deep Learning, Natural Language Processing and Image/Video Recognition techniques to different realms of application, such as smart automation, predictive engines, recommendation systems and conversational agents.