IBM is serving-up auto-generated micro-stories featuring its predictions on the probable outcomes of every tennis match scheduled for the U.S. Open this year.
Fed by data on each player — along with news media mentions about the players — IBM runs the info through its AI-powered super-computer Watson to come up with picks on winners and losers.
The result: Each micro-story features head-and-shoulders shots of the two players up top, followed by a pithy quote about each and closing with its short-and-sweet analysis on the match-up.
That breakdown is a number-cruncher’s delight, with supercomputer insights on tennis player Novak Djokovic, for example, like:
~Through the 1st Round, he (Djokovic) has a percentage of points with winners of 28%, 2nd among players in the field of 128
~Djokovic’s percentage of errors from the backhand is 56%, ranking 10th in the field
Meanwhile, Watson’s take on Djokovic’s opponent, Tallon Griekspoor, included:
~Through the 1st Round, his (Griekspoor’s) fastest serve speed is 149 MPH (240 KMH), 2nd among players in the field of 128
~Griekspoor’s percentage of points with backhand errors is 4%, versus 7% for Djokovic
All told, IBM’s auto-generated micro-stories are a formidable example of how AI can slice-and-dice the prospective outcome of a sports match in a way that is virtually impossible for even a team of human sports writers to equal — given the same amount of time to do their own analysis.
The U.S. Open runs through September 12 this year.
In other AI-generated writing news:
*In-Depth Review: Jarvis: Reviewer Jake Pfohl offers an extremely positive review on Jarvis in this piece — one of a number of auto-writers that use GPT-3 as their writing engine.
Like many GPT-3 based writers, Jarvis can auto-generate copy in seconds using scores of specific-use templates the software provides.
Observes Pfohl: “Overall, Jarvis.ai — formerly Conversion.ai — has taken the crown as the best AI copywriting tool on the market right now.
(Jarvis) offers over 50+ content templates for you to choose from, which makes it the perfect solution for any blogger or business owner that needs help with writing SEO-friendly blog posts, social media copywriting, ad campaigns, email subject lines, and more.”
This review is a good place to click if you want a detailed breakdown on virtually every facet of Jarvis, including its various pricing offerings.
Do yourself a favor, though, and also read another in-depth review of Jarvis, when it was marketed under the name Conversion.ai.
This review, written by an author with a masters degree, is also positive.
But it also offers more detail on the software’s shortcomings.
Other ‘quick study’ reviews on Jarvis are available on software review sites G2 Crowd and Capterra.
*New AI Auto-Writer Frees-Up Doctors to Interact With Patients: A new speech-to-text auto-writer from Nuance Communications is enabling docs to forgo taking notes on their encounters with patients.
Instead, the software — dubbed Dragon Ambient Experience — does all the notetaking by ‘listening in’ on the banter between doctor and patient during an appointment.
Subsequently, the tool’s AI contextualizes the interaction between the two and ultimately generates a clinical note on the meeting that the doctor can later review and sign.
Enabling doctors to eschew note-taking and writing during patient interactions, Dragon Ambient Experience helps doctors focus more on what first inspired them to get into medicine: Healing people.
Says Jared Pelo, a chief medical information officer at Nuance: “We’ve seen physician burn-out go from 72% to 17% after just a few months of using Nuance” Dragon Ambient Experience.
*Google Adds ‘Smart Reply’ to Google Docs Comments: Smart Reply — a feature found in Google Gmail that helps auto-generate a reply to an email — is now available as an auto-commenter for Google Docs users.
As with Gmail, Google Smart Reply analyzes a comment made in a collaborative Google document and then auto-generates a number of suggested replies.
Look for Smart Reply to roll-out to users of Google Docs in Google Workspace, G Suite Basic — as well as to users of Google business accounts and Google personal accounts.
*Conference: A Look at AI-Generated Writing’s Future: AI writing enthusiasts looking for the last word on the current state of automated writing — as well as its future — will want to check-out a definitive conference on the tech later this month.
Specifically: The 14th International Conference on Natural Language Generation, slated for September 20-24, in Aberdeen, United Kingdom.
It’s here that top researchers in AI-generated writing — also known as natural language generation — will offer-up their latest work and insights on a tech that is re-imagining how writing is done.
This year, much of the conference is happening online.
And all conference presentations of research work and posters will be online as well.
*A Look at ‘Old School’ AI Writing: This is a great article to check-out if you’re looking to get a true nuts-and-bolts look at how ‘old school’ AI writing works.
While a seemingly unending flurry of new auto-writers is hitting the market based on GPT-3 — a tech that relies on AI-driven predictive analytics — old school auto-writers leverage different methodology.
In a phrase, these commercial pioneers of AI writing work a lot like the Mad Libs game you probably played in grammar school.
Mad Labs furnished you a template story — say, how pizza was invented — you asked your friends for random nouns, verbs and adjectives to plug into the template and then everyone roared at the non sequiturs.
AI-generated reports work on the same principle.
They start out with the template of a say company report — a skeleton of a story, really — then automatically fleshed-out that story with data the AI software retrieves from your company database.
One of the major differences between Mad Libs and old school AI-generated writing is that each ‘blank’ in a company report can be the result of an extremely complex mathematical equation on say how company sales are faring — which is automatically computed by the AI software as it writes the company report.
Major AI-writing pioneering companies like Automated Insights, Arria NLG and AX Semantics use this method to auto-generate writing.
*New Tools in AI in Journalism: What’s Working: What’s New in Publishing offers an interesting look at some new AI tools news media outlets are using to stay competitive.
Key among those:
~Piano: A news curation tool that selects news stories to present to readers based on their reading habits and choice of reading device
~WhizzCo: Another news curation tool that suggests content a pub should buy, based on its ability to generate the most revenue for that publication
~GeoEdge: An AI-driven policing tool that quarantines ads it suspects are infected with malware and/or similar malicious code
*AI Writing in India: Doing Nicely, Thank You: Copyright spotlights a number of companies proliferating AI-generated writing across India in this article.
Worthy of note, according to Copyright, include:
~Phrazor: AI software that auto-generates company reports and similar text from organization databases
~Yantra: A bot that generates news in context
~Adziz: An auto article writer
*Brace for It: Mediocre Writing at Scale: Nate Nead is one of a number of AI watchers who fear auto-writers powered by GPT-3 will reduce the quality of writing across the Web.
Observes Nead, a writer for Readwrite: “There is a mediocre content deluge coming to the Internet the likes of which we have not seen.”
Currently, access to GPT-3 is closely guarded by maker OpenAI, which only licenses the tech to companies it trusts, according to Nead.
But Nead believes that firewall could be stormed in a heartbeat.
Observes Nead: “Keeping tight controls on a beta product (GPT-3) that could be used nefariously is more than smart.
“But it doesn’t mean would-be abusers won’t still find a way to circumvent the rules.”
(For an in-depth look at GPT-3, check out: “GPT-3 and AI Writing: Stunning, if Imperfect,” by Joe Dysart.)
*AI Big Picture: The U.S.’ Strategy: Law firm Gibson Dunn offers a quick study analysis on how the U.S. plans to optimize the use of
AI in this piece.
Included are eye-opening looks at:
~U.S. initiatives to foster AI development in country
~Fostering trust in AI systems
~AI and national security
Share a Link: Please consider sharing a link to https://RobotWritersAI.com from your blog, social media post, publication or emails. More links leading to RobotWritersAI.com helps everyone interested in AI-generated writing.
–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.