It appears we may not be ready to turn over important news curation duties to AI-driven editors, according to an opinion piece in Analytics India.
Case in point: Trusting AI to get the story right resulted in a major gaffe at AI-driven MSN News last week, observes Analytics India writer Ram Sugar.
The problem: MSN’s AI ran the wrong photo along with a piece on racism, which it curated from another online news source.
That triggered charges from the misidentified source — Jade Thirwall — that the AI software itself was plagued by racist programming.
Observes Sugar: “Having a completely automated information curator cannot be justified — unless some organization wants to hide behind the veil of AI by shifting the blame to a non-human entity.”
In other AI-generated writing news:
*Microsoft MSN’s Move to AI Editing: A Glimpse at the Future?: Microsoft’s decision to replace approximately 50 freelance editors with AI editors July 1 indicates more writing jobs will most likely be lost to AI, according to David Roe.
Roe is a writer for CMS Wire.
“Until now, it has been argued that the main jobs that will lose out to AI are unskilled, manual labor,” Roe observes. “However, the MSN decision indicates that this is not the case.”
Specifically, writers can expect to see jobs like proofreading disappear as AI continues to expand into journalism, copywriting, technical writing and similar genres of pro writing, according to Roe.
“Proofreaders are likely to be replaced by AI — despite the skill needed to be a great editor,” Roe observes. “An editor needs to have a good command of the English language.
“But many Web sites and companies are already using grammar check software like Hemingway App and Grammarly.
“There are plenty of technologies that make it easy to self-check your writing.”
*Partnering With AI to Write Sci-Fi: Stephen Marche, a writer with MIT Technology, says he has played with AI to write sci-fi – with impressive results.
“The resulting story—“Twinkle Twinkle,” published in Wired— not only looked and felt like a science fiction story,” Marche observes. “It also, to my surprise, contained an original narrative idea.”
Marche’s take: “Whether we want it or not, the machines are coming. The question is: how literature will respond.”
*OpenAI Releases Commercialized AI Text Generator: OpenAI is offering a commercialized version of the latest version of its AI text generator – GPT-3 – according to a story in Wired.
The cloud-based service works by taking an inputted sentence and then adding additional sentences it believes should immediately follow — based on its best, AI-aided guess.
While pre-commercial versions of the auto text generator were often used to fabricate all sorts of far-fetched musings — which sometimes resulted in nonsense text — OpenAI sees a number of practical applications for the commercial version.
For example: The system could be trained to auto-generate a list of companies, their stock symbols and foundation dates, according to Alex Hern, a writer for The Guardian.
That would work by first inputting the full data for a couple of companies, and then inputting just the names of other companies, Hern observes.
In a perfect world, the commercialized version of GPT-3 would sense a list was being requested and then auto-generate stock symbols and foundation dates for the additional companies.
Data would be gleaned from scans of the Web by GPT-3.
Founded as a nonprofit in 2015, OpenAI became a for-profit in 2018 and has been funded in part with a $1 billion investment from Microsoft, according to Tom Simonite, a writer for Wired.
*AI-Generated Emails: Will Authenticity Be Lost In Auto-Created Replies?: New research from Stanford University warns that Gmail’s AI-aided auto-replies to emails, such as “Sounds good!” “See you then!” or “Tuesday works for me!” may be inadvertently imbuing email replies with emotions that senders are not really feeling.
“The simple bias of being positive has implications at Google scale,” says Jeff Hancock, a professor of communication at Stanford University.
“Maybe it’s no big deal and it’s what people would have said anyway,” Hancock adds. “But we don’t know.”
Hancock also fears the increasing reliance on AI in everyday email could taint people’s trust in the messages they receive.
Plus, such auto-communications could even alter people’s relationships with one another, he adds.
“If people are uncertain and can’t tell if something is AI or human, then they are more suspicious,” Hancock says.
*Top Atlanta Newspaper Automating Real Estate Stories With AI: The Atlanta Journal Constitution is now forging articles on housing sales with AI-generated writing.
Essentially, the paper is using software from United Robots to drill down into county databases and produce short news stories on houses that sell — based on zip code.
A longtime player in AI-generated writing, United Robots was founded in 2015 and has automated news reporting for publishers in Sweden, Norway, Finland, Germany, Netherlands and Belgium.
*AI-Generated Writing Meets Metal Music: Music Web site Lyrics.rip has generated its own knock-off of a Metallica lyrics using artificial intelligence.
Specifically, the site fed the entire catalogue of Metallica’s lyrics into an AI-text generator and ‘let-it-rip’ to come up with its own lyrics.
The result: a song titled “Deliverance Rides,” with endearing verses like:
“Curses rest on our brains
No, there’s an evil feeling in vain
I can’t believe I stand alone
Beast under wicked sky”
Music accompaniment to the wide-eyed cyber-rants was written by a mere human.
*AI and Journalism: No Worries Here: David Tomchak, slated to become the head of digital and innovation at the UK’s New Statesman in September, is all-in on AI.
“Having read what AI’s can write — they can write really well,” Tomchak told the PressGazette. “You can write-up company results with it.
“It frees-up journalists who would otherwise be going through a whole bunch of company results and taking out the data to do other stuff that actually requires investigative skills.
“The whole idea is that machines make it easier,” Tomchak says. “They don’t take over.”
The New Statesman is a leading political and cultural magazine in the UK that journals through a progressive lens.
*Virtual Workshop: Preparing Your Newsroom for an AI-Powered Future: AI-generated news guru Charlie Beckett will be presenting a virtual workshop on how newsrooms can get the most from AI July 2.
The talk is part of a 4-day online conference, “Innovation in Digital Publishing During the Covid-19 Crisis,” squired by Newsrewired.
Beckett is director of the Media Policy Project at the London School of Economics and Political Science.
He’s also the driving force behind a group looking to prototype AI tools — JournalismAI.
Beckett also authored the study, “New powers, New Responsibilities. A global survey of journalism and artificial intelligence,” released last fall.
*Scientists Turn to AI to Sort Through Coronavirus Research: Programmers are currently designing a number of AI research tools that will enable Coronavirus researchers to quickly find the data they need in what has become a torrent of scientific findings on the virus, according to an article Nature.
While prototypes of some of the tools are currently available, their utility is largely unproven, according Nature writer Matthew Hutson.
Even so, developers hope the system will help researchers in some way.
*Special Feature: Company Reports That Write Themselves
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.