{{item.videoDuration}}
{{item.title}}
{{item.text}}
{{item.videoDuration}}
{{item.title}}
{{item.text}}
How Artificial Intelligence impacts Swiss Media and Telecom Industry
Artificial intelligence (AI) is currently experiencing a real boom. Even though the origins of the concept of AI go back to the 1950s, it is today that the concept is on the verge of a breakthrough. AI is the broader concept of machines being able to carry out tasks in a way that we refer to as “smart”. The latest breakthrough in AI is deep learning, a new machine learning method based on neural networks that learns and becomes even more accurate as we feed the model more data. This gives the impression that the machine is actually learning by itself. As the advance of AI gathers pace and scale, the Publishing and Telecom sector will be among the first to be transformed.
Use of Artificial Intelligence in Swiss media companies
In the publishing sector, there are already a number of applications for AI. Research algorithms, live monitoring, automatic translation, image recognition methods and even algorithms that write texts independently. For international publishers, AI is already an integral part of their business: The New York Times is using AI for data research in realtime; the Economist is reshaping its comments platform to foster thought-provoking, high-quality debate among its readers; and The Washington Post for example used AI for “automated storytelling” to make sport updates, tweets and blogs at the 2016 Rio Olympics.
Let’s take a look at Switzerland. How is AI used by the major media companies in this country? Does AI give the Swiss publishing industry a boost thanks to new monetization possibilities? Will the Swiss media companies, as in Asia, change into AI-companies in which more and more robots take over the functions of journalists? To answer these questions, we zeroed-in on three big Swiss media companies: Tamedia, Ringier and CH Media.
At all three of these companies, the potential of AI has already been recognised and innovative big AI solutions are in use. Ringier has developed a generic technology- & data-platform with AI applications and received the INMA "Global Media Award" for one of its corresponding use cases entitled "Using advanced artificial intelligence (AI) to boost digital reader engagement". Tamedia uses the research tool “Tadam” (Tamedia Data Mining), which also played a role in the unveiling of the Panama papers.
The Swiss media companies are currently developing and implementing various use cases for AI. The opportunities for future applications seems enormous. According to experts, AI offers Swiss media companies great potential in the areas of research, live monitoring and paid content models – not to mention robotic journalism in certain instances. Apart from that potential, though, limits are also seen in current AI solutions. So there is a big incentive for media companies to expand those limits through the further development of smart technologies.
“AI applications are cropping up faster than expected and will have a tremendous impact on the entire media business. AI is one of the global trends that are quickly making a breakthrough in the media industry. So we need to pay close attention to the evolution of AI.”
Do robots replace journalists and news agencies?
Leading media companies such as the Associated Press (AP), The New York Times and The Los Angeles Times already publish texts written by robots. Currently, these are short and less complex articles, for example standardised financial and sport reports.
Robo-journalism is also applied by Swiss media companies. For instance, at CH Media, the newly founded joint venture between AZ Medien and NZZ-Mediengruppe, robots will write e.g. stock market and company reports. A broad field of application for robo-journalism is sports reporting, where live tickers and sports reportage are increasingly written by robots. There are also video tools that allow the automated evaluation of football games as well as the instant production of highlight videos. Robo-journalism is often used for reporting on fringe sports, which otherwise would capture very little “ink” for reasons of efficiency.
Robots can replace journalists in certain steps of the content creation process. However, this does not eliminate the need for reporters. Robo-journalism relieves reporters and journalists of their everyday repetitive tasks, thereby enabling them to focus on investigative work and produce high-quality, in-depth reports. The demand for quality in paid content has risen enormously, since nowadays breaking news is available free of charge, everywhere. Publishers have to produce not just more content, but also better content. The only way they can remain competitive is by achieving broad reach while also delivering top-notch content.
"Media have a great impact on society and democracy, so well-researched and thoughtful journalism is very important. AI solutions represent an opportunity for the media industry and should be used for repetitive and time-consuming tasks. This makes it possible to invest the time and resources necessary for thoroughly researched, quality journalism."
AI has the ability to react instantaneously to real-time data and create the general outlines of a story. The risks of job cutbacks due to AI lie not so much with media companies but rather with news agencies. Editors can access media releases, sports tables and other sources directly and automatically and prepare the data-based content on their own. As a result, many middle-market retailers fall out of the value chain, the tasks are insourced and largely automated by the media companies.
Research algorithms and live monitoring are already well developed at the editorial offices of the Swiss media companies. The "Ringier TagCloud" – one of the component of the Sherlock platform - for example can automatically match appropriate images to contents, helps to create dossiers about a specific subject and automatically index content and websites for search engine optimisation (SEO) with an extremely high level of quality. These use cases reduce manual workload, enhance quality and improve efficiency.
Use case «Tamedia Data Mining – Tadam»
Tamedia has been developing the research tool “Tadam” (Tamedia Data Mining). Tadam is a platform that securely collects and stores massive amounts of unstructured data and makes it easily searchable, regardless of the format or language of the original source. The SonntagsZeitung and Le Matin Dimanche, using Tadam as a pilot, were contributors to the series of headline articles on the so-called Panama Papers published in 2016.
Tadam is also designed for daily use in editorials. It facilitates live monitoring tools that search the entire web and social networks as well as collect live messages from Swiss cantonal police stations. On average, journalists working with Tadam receive information 45 minutes earlier than they appear on the fastest news site. Tadam directly accesses various sources without waiting for second-hand reports from search engines or agencies. The tool has become indispensable at Tamedia especially for the sports desk and will be rolled out to other departments.
"To drive digital development with AI solutions, additional investments in methods, tools and corporate culture are crucial. Employees are involved in this research-process, with the result being that fears are reduced and no silos are created within the company."
How is AI driving paid content?
For years, publishers and brands alike have been creating content based on audience profiles. The challenge comes from the fact that the personalisation algorithms tend to display "more of the same" and the variety of topics is limited. Users end up in the so-called "filter bubble" and miss relevant news because they’re informed too one-sidedly. The consequence: consumers may turn a deaf ear or a blind eye to personalised offers.
Using intelligent algorithms, various factors are now being incorporated to make personal recommendations. These recommendations depend not only on what the user clicks, reads and buys, but also on the weather, location and current events. Smart personalised content has a surprise effect on the reader.
“To a certain degree, users appreciate personalised content that exactly matches their interests. However, if the surprise effect is missing from the recommendations, users tend to turn away and inform themselves elsewise.”
There is great potential for artificial intelligence in "AI paywalls". Previous paywall models were not tailored to the individual user and have therefore not achieved the desired effect. The Wall Street Journal as example is well advanced in dynamic monetisation. It has been building an AI paywall that adapts to reader behaviour and decides how many free (sample) articles they may access without signing up. The Wall Street Journal's paywall incorporates an intelligent algorithm that measures reader activity across 60 variables, including visit frequency, regionality, depth, favoured devices and preferred content types. This forms a propensity score, i.e. a person-specific subscription probability, which then helps to determine how many sample stories any given user can access. In short, reader activity shapes how much Wall Street Journal content they can read for free.
At CH Media AI is part of the paid content strategy and will be applied to the paywalls of the news portals. AI paywalls are expected to have great monetisation potential.
As for Tamedia, vast amounts of data are currently being evaluated in order to develop intelligent algorithms for monetisation models. The aim of AI paywalls is to record the individual reader's willingness to buy as early as possible. Some customers will even be offered premium content free of charge to make the product more attractive, while others will be asked to pay already the first time they view an article.
"AI's greatest potential for publishers lies in dynamic monetisation and personalization of the sales funnel. Static Paywalls as they are known today will no longer exist in 5 years. We need to address an individual user at the best point in time with a matching offering to improve significantly the conversion rates."
Use case «Ringier Sherlock»
As a part of Ringier’s ecosystem strategy, a generic technology and data platform was developed to apply in different types of use cases, across the entire group portfolio. What distinguishes the Ringier platform from other models is its generic nature and its large-scale, coordinated use of leading artificial intelligence technologies.
The platform enables several personalization and segmentation use cases. It generates fully automated content recommendations by directly linking relevant, up-to-date content found in other related articles (e.g. "More on this subject”). A cross-selling use case is “Affiliate Marketing”. Based on its in-depth grasp of interests and context, the platform can identify those users who are most likely to respond to an e-commerce product/service. These users are then shown personalized ads, thereby significantly increasing the click-through rate.
A number of further use cases are being developed in collaboration with the business units. Through the orchestration of AI technologies, the platform can be continuously updated and support various new business models. Ringier Sherlock was awarded the INMA Global Media Award in the category «Best idea to grow digital readership or engagement».
“The key uniqueness of this AI-powered technology- & data-platform is that it is generic and can be applied to many types of use cases. The purpose is to connect all the businesses within our group’s portfolio in order to leverage synergies and create a true business ecosystem.”
Does AI detect or trigger fake news?
When the topic turns to artificial intelligence, many critical voices are still to be heard. Smart technologies are suspect because they can be used to produce and distribute fake news and fake videos, i.e. to misinform and manipulate media consumers. The irony here is that artificial intelligence is supposed to recognise and neutralise precisely that type of misinformation.
The experts from the Swiss media companies we interviewed all believe that the use of AI poses no threat so long as there is a positive intention behind it. In terms of Switzerland, the danger of fake news is considered low. AI in and of itself is not evil; the trouble, though, is its inability to assess whether the input is reliable.
Without completely reliable data, a credibility gap opens up and the full potential of AI cannot be realised. The solution is to develop and implement mechanisms to ensure the authenticity of articles. The data sources must be screened with pattern recognition and fail-safes to filter-out malicious or misleading inputs.
AI requires large amounts of data to know what the correct response ought to be. And it takes huge quantities of data just to train the machine in the first place. Machines cannot make analogies the same way humans can. Face recognition, for example, requires billions of pictures before the machine can comprehend the patterns, learn from them and optimise/adapt itself accordingly. Human beings learn face recognition based on the reasonably manageable number of people in their environment. Without the availability of such data, AI’s capabilities are limited.
How advanced is AI in Swiss media compared to other countries?
The Swiss media industry can be considered progressive when it comes to the application of AI solutions. Compared to the USA or Asia, however, the dimensions are modest.
In Asia, the erstwhile media companies now label themselves as AI-companies. The mind set there is quite different. The media giants have morphed into technology companies and set their focus solely on user experience and personalisation as market competition in Asia is enormous. Without the aid of machines and AI, journalistic productions are no longer financeable. As in Asia, also in the US: personalised content is much more profitable than in Switzerland, since much greater economies of scale can be exploited. US newspapers serve international markets and their content is used all over the world.
Nonetheless, Europe is catching up as more and more crossborder interaction and networking has taken place amongst European countries. The Swiss media companies are focusing squarely on today’s global trends and participate actively in this exchange, be it by attending international trade fairs or partnering with major publishers such as The Washington Post or Shibsted. It is to be expected that Swiss media companies will make increasing use of AI in effort to heighten their efficiency and monetise their services – and these new monetisation models could represent the means for financing superior quality journalism in the future.
Artificial Intelligence in the Telecom Industry
As one of the fastest growing industrial sectors in the world, the telecommunications industry already has numerous applications for artificial intelligence.
Smart devices, including the smartphone, will be the vehicle through which AI will have the greatest impact on the telecom industry. One of the biggest implications of AI for telecoms will be the rise of conversational AI. In future the action required by the user will be triggered automatically by an AI interface, or otherwise by a conversational or gesture interface that interprets people’s everyday behaviours and habits. With facial recognition solutions built into mobile devices, it will be possible to customise marketing activities based on data collection. When the apps are activated, the user's facial expressions are captured and his/her state of mind is intuited or analysed to determine exactly what topic they are interested in. Emotional reactions to certain advertising spaces and videos will be tracked in order to tailor the advertising message to the consumer during the next campaign. Through facial recognition, the app will determine whether the commercial is really seen or whether the message is ignored and allow advertisers to book more accurate rates.
Today, Swiss telecom providers already use AI applications in certain ways. At Swisscom, various teams are working on the development and integration of artificial intelligences solutions.
“For Artificial Intelligence we see high potential in various fields of application. Swisscom is developing and integrating solutions to increase efficiency and process optimization in our customer service work and network operation but also to improve customer experience for example with the voice control function in Swiss German for Swisscom TV.”
Artificial intelligence is changing the way TV consumption takes place. With the use of AI algorithms, the television gets to know the users’ most popular type of programmes and only presents movies with the desired content. Today, voice recognition software is part of many TV platforms. Speech recognition technologies can be used to enable full control over the viewing experience without having to press a single button.
"Advanced data analytics and AI are essential for today’s TV offerings. We provide personalised content, enhanced with recommendations on the latest productions to keep users up-to-date. Also, AI-based voice assistants will be an integral part of future TV experiences."
AI also plays a major role in the development of 5G networks. A higher degree of automation is crucial to operating fifth-generation (5G) networks. The 5G mobile standard will lead to a rapid growth in data traffic and an increasing number of networked devices. This requires new concepts, such as cognitive – i.e. self-reactive and self-learning – monitoring systems.
{{filterContent.facetedTitle}}
{{contentList.dataService.numberHits}} {{contentList.dataService.numberHits == 1 ? 'result' : 'results'}}
{{contentList.loadingText}}
{{contentList.loadingText}}
Contact us
Bogdan Sutter
Director Advisory, Strategy und Transformation Expert, PwC Switzerland
Tel: +41 79 356 30 80
