Twenty Nineteen saw an embrace of AI-generated writing in virtually every major industry.
Increasing numbers of news outlets began to boldly experiment with the tools — as did advertisers and public relations pros.
Individual businesses also grabbed the technology to auto-write company reports based on info from their databases.
Plus, AI personalization – the ability to generate highly customized news and other writing based on what smartphone readers are tapping, swiping and commenting on as they interact with text – continued to crop-up in a number of implementations.
But the tech also inspired its fair share of controversy.
Many writers — especially journalists and copywriters — saw AI as synonymous with job loss.
And media watchers were alarmed as free AI software capable of generating fake news was released to the general public.
All told, scores of news stories chronicled the spread of AI-generated writing across the globe.
Here are the 12 that made the greatest impact:
*Newsrooms Eagerly Embracing AI-Generated Writing: Early successes with AI at major news outlets like the New York Times, the LA Times and Bloomberg spawned AI implementations in newsrooms worldwide.
China Science News, for example, is auto-generating science news based on article abstracts it finds in major scientific journals.
The Canadian Press brought in AI to automatically create instant background stories on specific electoral districts.
And the Houston Chronicle added AI news service Hoodline to auto-generate stories about high school sports, business openings and closings and weather.
Still other implementations representative of the spread of AI-generated writing across the globe included:
*The Professional Golfers Association’s use of AI to auto-generate 900+ golf stories every week
*United Robots’ deal to provide AI tools for Schibsted Media Group
*The BBC’s use of AI writing to generate 100 + stories every month – each hyper-localized for a specific community
*Writers Fearful AI May Spell Job Loss: Despite repeated assurances from publishers, many writers fear the AI looming in their doorways will be followed by pink slips.
Those concerns were reinforced by a new academic study, which found artificial intelligence generated writing and similar tools are currently being used to completely eliminate the jobs of human journalists at select news outlets.
Meanwhile, Mar Masson Mack, a writer at The Next Web, expressed unease over his AI robot ‘colleague,’ who worked alongside him, relentlessly punching out high-quality copy — sans any fatigue.
Overall writer anxiety about AI was best summarized by Thomas Kent, a journalism professor at Columbia University: “AI isn’t sentimental, and it isn’t ideological. It can be used for good or for evil. It can be used to replace journalists or to strengthen journalists.”
Adds Kent: “It’s a neutral piece of technology that can be adapted for almost anything.”
*Publishers Assuring Writers AI is an Aid –Not a Job Killer: Keenly aware that many writers gulp at the prospect of AI-generated writing, publishers released a flurry of articles in 2019 reassuring scribes that AI is their friend.
According to Nick Diakopoulos, for example, some estimates suggest that current levels of AI technology could automate only about 15% of a reporter’s job.
Diakopoulos is assistant professor, communication studies and computer science, Northwestern University.
Diakopoulos also says only 9% of an editor’s job can most likely be automated.
Plus, Diakopoulos also finds that more often than not, AI technologies appear to be creating new types of work in journalism — rather than eliminating jobs.
That sentiment was echoed by Molly Prosser, associate creative director at eBay. Prosser helped implement AI-generated writing of email subject heads at eBay.
“The hours I’ve spent editing and having team members pour over these things — it’s just meaningless work when we have a piece of AI that can do that for us,” Prosser says.
Still other stories assuring writers they’re safe from AI appeared in a major study released by the Media Policy Project, in the Indian Journalism Review and on Customer Think .
*AI Increasingly Used to Personalize Writing: AI has a unique ability to assess just what you’re looking to read – and deliver it to you, time and again.
News outlets have been among the first to capitalize on this advantage: “I think it’s going to be a crucial skill for every newsroom to stay up-to-date with those individual reader behaviors, and really learn from them,” says Jill Nicholson.
She’s head of product education at Chartbeat, a content analytics firm, who was quoted in a major analysis of AI’s use in the newsroom released by the European Broadcasters Union.
Meanwhile, eNewsletter service Rasa.io is personalizing newsletters by crawling the Web for content from hundreds of news sources. It auto-populates enewsletters with the select articles it finds, based on the preferences of the individual recipient.
But one of the most experienced of AI personalizers is UK News service Radar. It announced in September that it had personalized 200,000 stories for readers across Britain.
Even so, scholar Sophia Ignatidou warns that mainstream media’s new penchant for personalized news could undermine democracies.
“Telling audiences what they want – or expect – to hear is markedly different from telling them what they need to hear,” Ignatidou observes.
*Company Reports Are Writing Themselves: The same AI software newspapers are using to auto-generate stories is also being used to generate customized reports for businesses.
Essentially, these AI-generated writing solutions can drill-down into organization databases and auto-produce easy-to-understand, written reports based on that business data.
Anna Schena, for example, shared details of how her company adopts AI-generated writing to bring business databases to life.
And Sharon Daniels, CEO at Arria NLG — another AI-generated writing toolmaker — adds that businesses can look forward to Alexa-like, voice reports from their company databases.
*Advertisers Using AI for Short Snippets of Copy: While it appears AI still needs additional development before it can produce long-form, highly creative copy, it’s perfect for generating short snippets of advertising text.
Chase Bank, for example, turned-heads this past summer after it announced it was all-in on short ad slogans generated by artificial intelligence.
The bank inked a five-year deal with AI toolmaker Persado to create short ad slogans for its credit card and mortgage businesses.
Meanwhile, Textio – which specializes in creating job ads generated by AI – was recognized by CNBC as one of 50 private companies in 2019 “whose innovations are changing the world.”
And Health insurer Humana leveraged AI to enhance its email marketing.
Humana found that 31% of recipients opened an email it sent that sported a subject head created by AI.
That performance was 10% better than an email sent using a subject head created by a human.
*Significant Use of AI in Academic Publishing: Beyond the spotlight, academic publishers have been quietly onboarding robot writing tools for years, according to Publisher’s Weekly.
“Today, close to 80% of all our journal-content processing—from editorial services to final delivery—pass through various Integra products that are built on natural language processing and RPA (robotic process automation) frameworks,” says Sriram Subramanya, CEO, Integra Software Services.
Meanwhile, academic publisher Springer Nature released a prototype academic book written by AI.
*Public Relations Needs to Up-Its- Game With AI: While AI has made inroads in PR, too many PR firms are hanging back from the technology, according to Anne Gregory.
She’s a professor of corporate communications at University of Huddersfield.
“Other professions have already done major work on the shape of their future workforce — reviewing education and training, looking at their future role in organizations and society and at the ethics of AI,” Gregory observes.
“We need to get cracking, and get on with some serious work in all these areas.”
A study detailing — in part — public relations’ slow adaptation of AI, is due out January 2020.
*AI Becoming Standard in Consumer Editing Tools: Scores of AI editing tools are being released to help writers with grammar, spelling – and even the tone of their writing.
AI editing tool Grammarly, for example, now offers emoji’s that telegraph how your writing ‘feels’ to the reader.
Other AI editing tools for consumers include:
*Journalism Students Sorely Need AI Training: Given that AI appears destined to remake the face of journalism, j-students need to arrive at newsrooms with AI skills in tow, according to some journalism educators.
“We need to think about educational models to up these different literacies for journalists,” says Nick Diakopoulos. “We might imagine a doctorate of professional practice for computational journalism.”
Diakopoulos is assistant professor, communication studies and computer science, Northwestern University.
Jack Lule, chair, department of journalism and communication at Lehigh University, agrees: “We want to prepare students to not only be able to report on the transformations being wrought by AI, but also equip them to help shape its future for both society-at-large and the field of journalism.”
Cardiff University has already heeded this advice. Its students will be able to take master classes in AI journalism tools beginning February 2020.
“There are traditional print journalists who think ‘That’s not for me, I don’t need to think about AI’,” says Gavin Allen, a lecturer at Cardiff’s school of journalism.
“But it is becoming widely used at publications,” Allen adds. “AI is not the future — it is happening right now.
Apparently, many in the student population have gotten the word. This past fall, nearly 13,000 students from across the globe signed-up for a course in data journalism offered by the Knight Center for Journalism in the Americas.
“We were stunned at the enthusiasm for the course,” says Simon Rogers, one of the course instructors and data editor on Google’s News Lab team.
The six-week stint took students from 160 countries through “the entire pipeline of producing a data story,” Rogers says.
One of the most widely known applications of data journalism is the AI-generation of vivid graphs and other illustrations from info databases.
But data journalism also refers to the AI-generation of articles from databases.
*Fake News Generators Trigger a Flurry of Concern: The release of fake news generators like GPT-2 and Grover in 2019 triggered widespread uproar in the media world.
Many worry that such generators could easily spam fake news across social media and the Web – making it nearly impossible to discern what is real from what is propaganda.
“At the moment, the law has little to say about any of this,” says John Naughton.
He’s professor of the public understanding of technology at Open University. And he’s author of “From Gutenberg to Zuckerberg: What You Really Need to Know About the Internet.”
“We’re currently at the same stage as we were when governments first started thinking about regulating medicinal drugs,” Naughton adds.
Major tech companies like Apple, Amazon, Google, Facebook, IBM, and Microsoft are so concerned about fake news, they’ve joined a coalition to develop tools to fight the scourge.
“The ability to create synthetic or manipulated content that is difficult to discern from real events frames the urgent need for developing new capabilities for detecting such content,” observes Terah Lyons, executive director, The Partnership on AI.
But other media watchers warn that fear over fake news tech — at least fake news generators like GPT-2 — may be overblown.
Joshua Benton, director, Niemnan Journalism Lab, sees the perceived threat of GPT-2, for example, as more of a yawn.
*Gimmicky AI Writing Tricks Muddy Perception of What’s Possible: Given the nature of AI-generated writing tools, its no surprise people are hooking-up AI to outlandish ideas and pressing “Go.”
Granted, reports of these experiments often make for amusing brain candy.
But in the end, all they really do is muddy the perception of what’s possible with AI-generated writing.
For example: The Web was riddled with reports in 2019 and the year prior that AI had ‘written its first novel.’
But closer inspection of the stunt revealed that an AI system simply had been strapped into vehicle and driven across U.S. – continually fed along the way with data generated by a microphone, camera and GPS system.
The result: A book-length bunch of erratic, often nonsensical sentences that featured no plot, no point-of-view and no inner logic.
Essentially, the resulting ‘work’ is not a ‘novel’ at all.
And to describe it as such makes reading about the implementation a ridiculous waste of time.
Similarly, a new sci-fi ‘bookstore’ has popped-up on the Web, featuring ‘fiction’ generated by AI.
In reality, the ‘fiction’ created by the tech again makes no sense — it’s just gibberish.
Even the on site ‘reviewers’ for the ‘bookstore’ are creations of AI.
This time around, the AI experimenter behind the project was clear to detail the limitations of his ‘books.’
But again, it’s dangerous to call a piece of writing ‘fiction’ if in reality, it’s simply a collection of sentences that when taken together, are nonsense.
Feel free to send a link to RobotWritersAI.com to a friend or colleague.
*Also on RobotWritersAI.com — Evergreen Article:
*AI-Created Newsletters: On The Cheap
–Joe Dysart is editor of RobotWritersAI.com and a tech journalist with 20+ years experience. His work has appeared in 150+ publications, including The New York Times and the Financial Times of London.