AI Forged Opinion

AI Writing’s Next Leap: Robots That Understand

While more than a few people are wowed by the ability of artificial intelligence to auto-generate short stories, the tech has a ways to go before it can create deeply thoughtful works, according to Ankur A. Patel, vice president, Data Science, 7Park Data.

#ad

“AI does not truly understand language,” Patel observes. “To evolve materially, AI will need to truly understand the meaning of language to form coherent arguments and thoughts as it writes.”

Even so, AI-generated writing in its current state is very good at producing simple stories, according to Patel.

And “we should expect more of the writing we do — especially the more basic variant — to come from AI,” he adds.

At 7ParkData, Ankur and his data science team use alternative data to build data products for hedge funds and corporations.

They also develop machine learning as a service for enterprise clients.

Deeply familiar with the way AI works, Patel is also author of the new book, “Hands-On Unsupervised Learning Using Python.”

It’s a text on how to build machine learning solutions.

In other AI-generated writing news:

*Associated Press Expands Soccer Coverage With AI-Generated Writing: A pioneer in AI-generated writing, AP is expanding its use of the tech to offer automated previews of major league soccer games.

“With growing interest in the MLS (Major League Soccer) nationwide, we’re glad we can now provide insightful previews through this collaboration with Data Skrive and Sportradar,” says Barry Bedlan, AP’s director of vertical products.

The automation technology enables AP to generate the previews directly from data — without the intervention of editors, according to Bedlan.

*Grammarly’s Secret Sauce: Doing One Thing, Very Well: AI editing service Grammarly – which recently snagged $90 million in new funding – has cracked-the-code as an AI application, according to author Ben Dickson.

The package’s secret: Homing-in on a very specific task to ensure your grammar sparkles.

“In my experience using Grammarly, I’ve agreed with most of the corrections it gives me, and only ignore the ones that have more to do with style and personal preferences,” observes author Ben Dickson.

“Grammarly’s minimalist interface also makes sure that it does not venture out of the domain of narrow AI,” according to Dickson.

“Unlike AI assistants, Grammarly doesn’t have a human name, it doesn’t try to speak in casual human prose, and it has no voice interface. It’s barebones grammar suggestions. There’s no room for misunderstanding or bloated expectations.”

In a phrase, Grammarly is “the sweet spot of deep learning and natural language processing,” Dickson observes.

*Computers: Now Better ReadersThan Mere Humans?: BERT, a machine reading tool created by Google, can now score higher on reading comprehension tests than everyday mortals, according to Wired Magazine.

Even so, computer scientists and other academics question if BERT truly understands what it’s reading.

Essentially, the researchers wonder if tools like BERT are “really understanding language,” observes Tal Linzen, a computational linguist at Johns Hopkins University.

In fact, Linzen wonders if BERT has found a way to pick-up on “weird tricks that happen to work on the data sets that we commonly evaluate our models on.”

*Major Player in AI Journalism Releasing New Book in 2020: Francesco Marconi — who led development of AI-generated writing and similar practices at the Associated Press and the Wall Street Journal — is coming out with his perspective on AI in January.

The tome, “Newsmakers: Artificial Intelligence and the Future of Journalism,” promises to show newsrooms of all sizes how to use AI tools to tell news stories and connect with readers in new ways.

*Promoting Diversity? Show Me Your Words: Corporations that regularly sing-the-praises of diversity often come-up short on that ideal in their job ads.

“Textio has found that many of these companies still exhibit strong linguistic bias in recruiting and hiring content,” observes Julie Yue, data insights manager, Textio.

“Some use language that hardly sounds inclusive,” Yue adds. “And only a few have even articulated why diversity is good for their business.”

Not surprisingly, Yue indicates the solution to the diversity problem in job ads is to use an AI tool like Textio.

It’s AI software that generates job ads reflecting and promoting the value of diversity.

*AI-Generated Journalism in Action: Radar, an upstart AI-driven news service in the UK, has put together an insightful sampling of the kinds of stories it generates with AI tools.

The UK-based service is squired by six journalists. They regularly create story templates, which are used by AI to generate highly localized stories – based on data from government and other databases.

Among the hyper-local stories and trends Radar has unearthed:

*Hackney: The home of bike theft in London

*Figures reveal NHS spends millions prescribing gluten-free food for patients

*Britons spend L900 million-a-year charging electrical gadgets

*Lenovo Uses AI-Generated Writing to Up-Skill Chinese Writers: The technology giant was able to significantly improve the writing skills of its writing team by bringing in an AI-generated writing solution from Acrolinx, according to Keith Vicek, a senior software engineer at Lenovo.

Specifically, the AI tool gave the Chinese writers instant feedback and suggestions to improve they’re writing as they attempted to author documents in English.

All told, Acrolinx’s software enabled the writers to produce better documents on-the-spot – as well as become better writers in English over time, Vicek says.

*Deloitte Generating Internal Company Reports With AI:
The Big Four accounting services titan has onboarded AI to help with report writing.

Specifically, Deloitte uses AI-generated writing to auto-produce risk and compliance reports with AI, according to Mario Berger, a director at Deloitte.

“My focus is on reporting and regulatory reporting, the implementation of regulatory requirements for reporting and the digitization of all these topics,” Berger observes.

“I also moderate various formats such as strategy, innovation and business workshops and manage projects of varying scope and content,” he adds.

Berger says he chose AX Semantics’ tool for Deloitte’s AI-generated writing: “Quality and applicability of the solution are very high, which was also confirmed by the Forrester report from October 2018. “

“And last but not least, the operational cooperation between the teams is excellent,” Berger says. “We get-on really well with the AX Management Team.”

*Free Course in Data Journalism and Visualization: The Knight Center for Journalism in the Americas is offering a free course on how to use free AI tools for data journalism.

Data journalists often use AI-powered tools to auto-generate charts, graphics, info-graphics and similar visualizations from databases.

*Also on RobotWritersAI.com — Evergreen Article:

*AI-Created Newsletters: On The Cheap

Grammarly
#ad

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.

Never Miss An Issue
Join our newsletter to be instantly updated when the latest issue of Robot Writers AI publishes
We respect your privacy. Unsubscribe at any time -- we abhor spam as much as you do.
Subscribe
Please notify me when new comments are posted about this article.
guest

0 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments