I ROBOT

© Andrea Danti thinkstock

Artificial intelligence and machine learning tools are making significant inroads into publishing processes and product development.

Those of us who may be short on time and haven’t been able to get to that autobiography we’ve been meaning to write need worry no longer:  Artillect Publishing will do the work for you by scanning your online presence, merging some ancillary information and producing your sure-to-be best-selling biography.  Using artificial intelligence in its production process, Artillect is just one example of the increasing number of applications of artificial intelligence (AI) replacing inefficient processes, creating new products and adding insight to publishing.

News organizations including the Washington Post and Associated Press have been using artificial intelligence tools to create news reports for weather, sports and financial reporting where interpretation of the day’s (or game’s) activity can be fairly straightforward.   As these uses have grown in acceptance and utility, the use of AI to deliver more complex products is also growing.  AI tools can analyze text, images and data and deliver to a journalist sufficiently structured content around which they can build articles and stories.  AI tools can do this faster, more comprehensively and with greater accuracy than traditional research methods.  In analyzing text or images, AI tools can characterize the content: Positive/negative, liberal/conservative, for example.   Journalists have even used this capability to change editorial content to match specific political viewpoints, creating liberal, center and/or conservative versions of the same article.

The accuracy (or veracity) of news content generally sparks emotions and Knowwhere is applying AI and ML to the detection of ‘fakenews’.  Their process aggregates stories from more than a thousand different sources of varying political persuasions to create a “knowledge graph” or database of each news story which is then reviewed by (human) editors.  Mediashift recently reported that USAToday uses AI to create visual and interactive content, which they refer to as ‘dynamic editorial’ and hope will improve attention and raise reader engagement.

Journalism is by no means the only focus for AI and many companies across the media landscape are using and experimenting with AI tools.  In the last few days, for example, journal publisher Taylor & Francis (T&F) announced two partnerships with AI companies to add AI tools to their editorial processes.   In the first of these, T&F are working with Katalyst Technologies to create “contextual copyediting” using AI and natural language processing to assess and score the language quality of articles accepted into their journal’s workflow.  This use of AI is designed to make their editorial process more efficient by identifying and classifying journal submissions.  In the second example, the company announced they are working with UNSILO which will help T&F subscribers optimize the use of T&F content by surfacing additional, highly relevant content based on what they are already reading.   T&F expects that this function will enable users to discover new and pertinent research.

Next month, at the Society of Scholarly Publishers (SSP) meeting in Chicago, I am hosting a panel discussion on how publishers are using AI and machine learning (ML) to rethink how they manage their businesses.  On the panel, I will have executives from Storyfit, Yewno, Unsilo and Molecular Connections.   In advance of that meeting, I thought it would be interesting to explore some of the companies offering their versions of AI to publishers:

StoryFit: Uses machine learning and data analysis to predict content marketability, improve discovery and drive sales for publishers and movie studios.

Yewno: Uses machine-learning and computational linguistics analyze high-quality content to extract concepts and discern patterns and relationships to make large volumes of information more effective.

UNSILO: Provides artificial intelligence tools and solutions for publishers to grow new business opportunities and improve existing customer experiences and workflows.

Bibblio: A recommender system that helps content businesses and publishers deliver more relevant, engaging discovery experiences to their users.

Iris.ai: Helps researchers build precise reading lists of research documents across three parameters: Key information extraction (marking of possible contextually disambiguating information, forming basis for a document fingerprint); Neural Topic Modeling (clustering of semantically similar documents, cluster labeling, document fingerprint update); WISDM (basis for document fingerprint indexing. Fingerprints are matched using the WISDM document similarity metric.)

ECHOBOX: Helps content owners automate their social media presence and “is the best social media platform for publishers.”   Their solution is the “first artificial intelligence that increases your reach.”

Copyleaks:  Detects plagiarized content in scholarly workflows.

StatReviewer:  Uses ML to generate statistical and methodological reviews for scientific manuscripts to discern the integrity for scientific manuscripts.  Product is currently in private beta.

Webbitz: An AI-powered video creation platform that leverages patented text-to-video technology to streamline production of original short-form videos across multiple platforms.  The company works with more than 400 publishers.

Ross Intelligence: Using a combination of IBM Watson and proprietary algorithms, ROSS is the AI-driven successor to tools like LexisNexis and supports both legal discovery and legal research findings.

ScriptBook: Employs artificial intelligence to analyze screenplays.  The Script2Screen-solution delivers an objective assessment of a script's commercial and critical success.  The company believes their AI-driven process delivers both objective and fair results.  It treats every screenplay equally with no bias.

Arkadium: An AI product company which has been used by Sports Illustrated to create infographics.

There is little doubt that AI and ML applications and tools are here to stay within publishing workflows.  But it’s also true that we are only at the beginning of what will be a fundamental shift in workforce management, application and roles.   The prevailing concern is that machines will take all the jobs and leave humans with nothing left to do.  Not only is this excessive hyperbole, the impact of technology on work activities over the last 200 years points more towards a shift in the role of human work rather than the total replacement of workers.   When ATMs were first rolled out more than 30 years ago, there was an expectation that brick and mortar banks and their tellers would disappear.  In fact, the opposite occurred and teller employment increased as banks focused on building closer customer relationships.

AI will improve publisher workflows but will also help expand the utility and benefits of the products we are producing for customers.  Many of the newspapers mentioned above have been able to materially broaden their reporting in a way which would not have been possible without the type of functionality that AI offers.  Expect to see many more examples of this technology within the publishing business more broadly.