17 great quotes about AI limitations

/

Machine thinking is great for understanding the behavioral patterns across populations. It is not great for understanding the unique individual right in front of you. If you can understand another person’s perspective, you have a more valuable skill than the skill possessed by some machine vacuuming up vast masses of data about no one in particular. New York Times

Ian Bogost suggests that ChatGPT produces “an icon of the answer … rather than the answer itself.”  The Atlantic

A large language model is not capable of conducting independent research or gathering new information. It is only capable of generating text based on the input it is given, so it would not be able to provide original insights or perspectives on the topic at hand. Inside Higher Ed

The ability to create and give a good speech, connect with an audience, and organize fun and productive gatherings seem like a suite of skills that A.I. will not replicate. New York Times

The idea that “AI” can navigate contested terrain by flagging “disagreement” and synthesizing links to “both sides” is hardly sufficient. Such illusions of balance obscure the need to situate information and differentiate among sources: precisely the critical skills that college writing was designed to cultivate and empower. Public Books 

Something I noticed when I asked ChatGPT to write a short story: It makes everything sound like an unfunny parody. New York Magazine

I’ve learned that it is being used for such daily tasks as: translating code from one programming language to another, potentially saving hours spent searching web forums for a solution; generating plain-language summaries of published research, or identifying key arguments on a particular topic; and creating bullet points to pull into a presentation or lecture. Chronicle of Higher Ed

If AI-generated forensic sketches are ever released to the public, they can reinforce stereotypes and racial biases and can hamper an investigation by directing attention to people who look like the sketch instead of the actual perpetrator. Vice

AI feels mundane. It just feels like using any other technology. So we really need to reckon with our own expectations, turn down the hype, and close the gap between what we imagine and what the reality is. The Markup

The information produced by AI language models and chatbots is often incorrect. The tricky thing is that when it’s wrong, it’s wrong in ways that are difficult to spot. The Verge

Our tests found that it sometimes offers responses that potentially include plagiarism, contradict itself, are factually incorrect or have grammatical errors, to name a few — all of which could be problematic at work. Washington Post

When we discuss hallucinations and out-of-date databases, we should be careful about reaching summative judgments. These products are still very much in development; there will be new innovations, and there will be bigger and better pools of data that will stir the pot among ranking brands and products. Inside Higher Ed

CNET started quietly publishing articles explaining financial topics using “‘automated technology’ – a stylistic euphemism for AI,” CNET had to issue corrections on 41 of the 77 stories after uncovering errors despite the articles being reviewed by humans prior to publication. Some of the errors came down to basic math. Columbia Journalism Review

I think the questionable accuracy of responses provided by ChatGPT is its biggest downside. It means the user is responsible for verifying the information, which takes away the ease people are attributing to ChatGPT. Demand Sage

ChatGPT has proven inept at reproducing even the simplest ideas in rocketry. In addition to messing up the rocket equation, it bungled concepts such as the thrust-to-weight ratio, a basic measure of the rocket's ability to fly. NPR

ChatGPT can write poemlike streams of regurgitated text, but . . . they don’t satisfy the minimal criterion of a poem, which is a pattern of language that compresses the messy data of experience, emotion, truth, or knowledge and turns those, as W. H. Auden wrote in 1935, into “memorable speech.” The Atlantic

Even if researchers trained these systems solely on peer-reviewed scientific literature, they might still produce statements that were scientifically ridiculous. Even if they learned solely from text that was true, they might still produce untruths. Even if they learned only from text that was wholesome, they might still generate something creepy. New York Times

8 AI Predictions

/

Humans will specialize in whatever AI does worst. Chronicle of Higher Ed

Many types of work will be taken over by machines, and jobs will vanish. This change is typically seen as a cause for gloom. I suggest we see it as an opportunity to revitalize education by replacing unsatisfying work with meaningful labor. Chronicle of Higher Ed

AI will certainly force us to concentrate on those talents and skills that will remain uniquely human. Chronicle of Higher Ed 

Will writers start proclaiming they are “natural” writers, with no AI use in their work, akin to bodybuilders who choose not to use performance-enhancing drugs? Washington Post

It’s going to creep into our lives in ways we least expect it  Wall Street Journal

The role of software engineers will evolve into one of guiding and overseeing the AI's work, providing input and feedback, and ensuring that the generated code meets the project's requirements.  Prompt engineering will be critical in using automated code generators as prompts must be carefully crafted to accurately capture the intent of the desired code.  Forbes

While I think that A.I. tools help express our creativity, creativity will still be the driving force behind the future of art. New York Magazine

The new web is struggling to be born, and the decisions we make now will shape how it grows  The Verge  

7 Webinars in the next 9 days about AI, teaching, media law, broadcast, journalism & social media

/

Thu, July 6 - Ethics and Privacy in the Age of AI

What: Our speakers will introduce you to the foundations of data ethics and privacy, how they intersect, and how to evaluate ethical, privacy, and legal principles if you choose to leverage generative AI tools in your fundraising and engagement practices.

Who: Carrie Cobb, VP, Chief Data Officer at Blackbaud; Cameron Stoll, Sr. Director of Privacy Management at Blackbaud

When: 12 noon, Eastern

Where: Zoom

Cost: Free

Sponsor: Blackbaud

More Info

 

Thu, July 6 - The Future of Education in the Age of AI: Trends, Issues and Solutions

What: This workshop explores current and future trends, issues, and solutions of AI in education, emphasizing human-in-the-loop (HITL) approaches to engage educators and learners in design, development, and evaluation. Participants will explore the latest research and best practices to critically and creatively use AI to improve teaching and learning. They will also get an opportunity to network with other researchers and practitioners interested in AI and education.

Who: W. Ian O'Byrne Associate Professor of Literacy Education, College of Charleston

When: 12 noon, Eastern

Where: Zoom

Cost: Free

Sponsor: Media Education Lab

More Info

 

Mon, July 10 - Media Law Office Hours

What: Journalists with legal questions to help find answers with an attorney who specializes in this area.  

Who: Attorney Matthew Leish

When: 5 pm, Eastern

Where: Zoom

Cost: Free

Sponsor: New York’s Deadline Club

More info

 

Tue, July 11 - A Practical Newsroom Guide to Artificial Intelligence

What: We’ll lift up some examples of AI being put to use, share guidelines for implementing it and talk about approachable ways of testing this in your newsroom.

Who: Aimee Rinehart, Senior Product Manager AI Strategy at The Associated Press, will moderate this discussion; Joe Amditis, Assistant Director of Products and Events, Center for Cooperative Media; Ernest Kung, AI Product Manager, The Associated Press; Adriana Lacy, Adriana Lacy Consulting (2021 MJ Bear Fellow); Hank Sims, Editor and Lead Developer, Lost Coast Outpost

When: 12 noon, Eastern

Where: Zoom

Cost: Free for ONA members ($25 for nonmembers)

Sponsor: Online News Association

More Info

 

Tue, July 11 - Social Media 101 for Nonprofits

What: Practical tips and tools for extending your cause and mission via social media. We cover the basics of using social media for your nonprofit organization and give you handy tips for the “big 3:” Facebook, Twitter and LinkedIn.

Who: Kiersten Hill, Director of Nonprofit Solutions

When: 2 pm, Central

Where: Zoom

Cost: Free

Sponsor: FireSpring

More Info

 

Wed, July 12 - Digging Deeper in Broadcast Journalism

What: This lesson will focus on how anyone involved in gathering information for broadcast news can use investigative techniques in his or her daily work despite the constraints of busy newsrooms and deadline pressures. You’ll learn:  How to incorporate data into your stories (and convince your editor to let you do so). How to use public records to strengthen your reporting. How to find exclusives and best pitch them to your manager.

Who: Kevin Rothstein is a producer with the investigative unit at WCVB-TV in Boston, specializing in covering criminal justice issues and vulnerable populations.

When: Noon, Eastern

Where: Zoom

Cost: Free

Sponsor: International Center for Journalists

More Info

 

Friday, July 14 - Elevate your journalism by collaborating

What: This session will examine the benefits and even a few pitfalls of news collaborations. Plus, you’ll see examples of success stories to spur you to look for different ways to collaborate with other newsrooms and organizations to tell the most important stories to your readers.

Who: Jean Hodges of Gannett

When: 1 pm, Central

Where: Zoom

Cost: Free

Sponsor: West Virginia Press Association

More Info 

The new web struggles to be born

/

The changes AI is currently causing are just the latest in a long struggle in the web’s history. Essentially, this is a battle over information — over who makes it, how you access it, and who gets paid. But just because the fight is familiar doesn’t mean it doesn’t matter, nor does it guarantee the system that follows will be better than what we have now. The new web is struggling to be born, and the decisions we make now will shape how it grows.

James Vincent writing in The Verge

27 Data Science & AI articles from June 2023

/

An argument for bigger quantum neural networks

In-orbit demonstration of a re-trainable machine learning payload for processing optical imagery

7 ways ChatGPT makes you code better and faster

Are data scientists still needed in the age of generative AI? Not according to this opinion piece

Making Predictions: A Beginner’s Guide to Linear Regression in Python

Air Force studying ‘military applications’ for artificial intelligence like ChatGPT  

24 articles worth reading about the dangers of AI (beyond security issues)

Open-source AI chatbots are booming

A hacking conference (DEF CON 31) has invited hackers to find bugs and biases in AI  

9 articles worth reading about the security dangers of AI

Neural Networks need data even fake data to learn: Why researchers turn to synthetic data to train their artificial intelligence systems

China tests first-ever low-Earth orbit constellation to rival SpaceX's Starlink

Intelligence analysts confront the reality of deepfakes  

The NGA is hailing the value of AI tools & machine learning to analyze 1000s of satellite images

How the rise of low Earth orbit satellites can disrupt how militaries fight

Space Force reconsiders the use of the Global Positioning System constellation

USGIF white paper on GEOINT opportunities created by AI related to synthetic training data

A look at how the commercial satellite economy got to where it is today

Mastering the art of data storytelling: A guide for data scientists

A system based on Google DeepMind’s AlphaZero AI can create algorithms that will sort data faster than algorithms built by people

A visual introduction to neural networks

Asking ChatGPT to write you a malicious code

Mutating malware can be built using the ChatGPT

New US spy satellites to track Chinese, Russian threats in orbit

Five ways to help your data science team collaborate more effectively

Many commercial-satellite operators are still creating overly ambitious plans

NGA: AI has come a long way but “not good enough” to justify a pause in development

Don’t make waves

/

Thomas Moriarty arranged an experiment in which innocent persons would be practically accused of stealing. The experimental aide would stand behind an adult businessman making a call in a phone booth in Grand Central Station; when the call was completed the aid would play out the following script: “Excuse me, I was here a few minutes ago I left my ring on the counter under the phone. Did you find it?” Of course, all subjects replied, “No.”

The aide would then say, “I've got to find it. Are you sure you didn't see it? Sometimes people pick things up without thinking about it. Again, subjects would deny having seen the ring. Then the aid would ask, “Would you empty your pockets?”

The Investigators wondered how many people would comply with such an overbearing request, one which amounts to an allegation of petty thievery. The compliance rate was 80 percent: four of every five adult males essentially submitted to a search by emptying their pockets. The percentages were even higher in laboratory experiments. And even when a “disinterested bystander” said to the aide. “You’ve got no right to ask him to empty his pockets,” the subjects still complied.

Such studies show how prevalent passivity is. It is alarming that so few people are willing to stand up for their rights when they are being put upon and clearly annoyed. Apparently, most of us would rather not get into a hassle about anything, especially with a stranger. The slogan is: Don’t make waves.

Sharon and Gordon Bower, Asserting Yourself

June posts about AI

/

Making exterior changes to avoid internal changes

/

Some people make changes so they won’t have to make transitions. They walk out on their marriages, but take along the attitudes toward partners that destroyed their marriages. Or they continue to search for “someone to take care of me” after they quit their jobs because their bosses are not interested in playing that role. Or they move because their town doesn’t have any “interesting people” in it—only to find that their new town doesn’t either. Such people may claim that they are “always in transition,” but in fact they are probably never in transition. They are addicted to change, and like any addiction, it is an escape from the real issues raised by their lives.

William Bridges, The Way of Transition

18 wild things people are trying to get AI to do

/

Can AI read my mind?

New AI system could help people who lost their ability to speak - CBS News

Can AI rap?

A Swedish newspaper is having AI rap its articles in an attempt to get young people interested in the news - Business Insider

Are AI pets available?

AI pets are booming: They can include realistic programmed personalities — plus tails that wag - AI Time Journal  

Can AI past an MBA test?

ChatGPT passes Wharton Business School's MBA exam, gets a B - Interesting Engineering

Can AI simulate large-scale economic or political events?

Generative AI landscape: Potential future trends - Tech Target

Can AI contest parking tickets?

I asked ChatGPT to contest my parking ticket - Fast Company

Can AI set insurance rates?

Generative AI is helping figure out who is riskier for insurers – Semafor

Can AI create fashion?

Generative AI: Unlocking the future of fashion - McKinsey

Can AI explain history?

How AI is helping historians better understand our past - MIT Tech Review 

Can an AI be my lawyer’s assistant?

Why it’s imperative lawyers adopt a ‘legal copilot’ model with AI – Legal Dive

LexisNexis has launched a generative AI tool that can draft documents, conduct research & summarize legal issues- LawNext  

Can AI make decent movies?

Another Reason Hollywood Will Love AI - Wall Street Journal

Welcome to the new surreal. How AI-generated video is changing film - MIT Tech Review

Can AI spot materials inside of images?

Researchers use AI to identify similar materials in images - MIT Tech Review

Can AI pick hit songs?

Accurately predicting hit songs using neurophysiology and machine learning - Frontiers

Neuro-forecasting the next No. 1 song - Axios

Can AI replace data scientists?

Are data scientists still needed in the age of generative AI? - KD Nuggets

Can AI plan your trip better than you can?

In Milan, Putting an A.I. Travel Adviser to the Test - New York Times 

Can AI build a website?

How to use AI Art and ChatGPT to Create Insane Web Designs - Codex Community (video)

Can AI play Minecraft?

They Plugged GPT-4 Into Minecraft—and Unearthed New Potential for AI - Wired

Can AI provide commentary at tennis matches?

Wimbledon to introduce AI-powered commentary to coverage this year – The Guardian

Manufacturing unhappiness

/

While riding a train, a minister observed a handsome elderly couple. The wife was dressed in furs and diamonds and was the envy of all the other women on the train, but her disposition did not match her appearance. She constantly complained about everything. She criticized the food, the service, and the railway car.

Her husband, who was striking in appearance, was a calm, considerate man. As his wife raved about the abominable service, he sat calmly, occasionally showing embarrassment at his wife’s attitude. In an attempt to change the tone of the conversation, the husband engaged the minister in a conversation in which they discussed their occupations. Finally the husband said with a grin on his face, “My wife is in the manufacturing business.”

This surprised the minister because the wife certainly did not seem to be the executive type. The minster asked, “What does your wife manufacture?”

“Unhappiness,” her husband replied. “She manufactures her own unhappiness.”

Larry Kennedy, Down With Anxiety

10 questions about AI and medicine

/

Can AI diagnose illnesses?

The AI Will See You Now - Wall Street Journal

Can AI spot lung cancer years in advance?

AI tool could help spot lung cancer years in advance – Washington Post

Can AI answer your medical questions?

ChatGPT Will See You Now: Doctors Using AI to Answer Patient Questions - Wall Street Journal

ChatGPT improves their ability to communicate empathetically with patients – New York Times

Can AI write medical papers?

A Doctor Published Several Research Papers With Breakneck Speed. ChatGPT Wrote Them All - Digg

Can AI save someone’s voice?

Patients were told their voices could disappear. They turned to AI to save them - Washington Post  

Can AI predict medical outcomes?

The algorithm has been trained to make medical predictions based on reading genomes - Washington Post  

Can AI find new medicines?

Scientists have used AI to discover a new antibiotic that can kill a deadly species of superbug - BBC

Can AI spot cancer?

AI Tool Assists in Predicting the Likelihood of Pancreatic Cancer - Healthy Analytics

What role will AI play in medicine?

For now, the new AI in health care is going to be less a genius partner than a tireless scribe - New York Times

What limitations are we seeing for AI operating in the medical space?

An AI Diagnosed Her with Breast Cancer. Then She Ran an Experiment to See How Accurate It Was - The Markup

From “Dunce” to Genius

/

When Victor Serebriakoff was fifteen, his teacher told him he would never finish school and that he should drop out and learn a trade. Victor took the advice and for the next seventeen years he was an itinerant doing a variety of odd jobs. He had been told he was a "dunce" and for seventeen years he acted like one. When he was 32 years old, an amazing transformation took place. An evaluation revealed that he was a genius with an I.Q. of 161. Guess what? That's right, he started acting like a genius. Since that time he has written books, secured a number of patents and has become a successful businessman. Perhaps the most significant event for the former dropout was his election as chairman of the International Mensa Society. The Mensa Society has only one membership qualification, an I.Q. of 140.

The story of Victor Serebriakoff makes you wonder how many geniuses we have wandering around acting like dunces because someone told them they weren't too bright. Obviously, Victor did not suddenly acquire a tremendous amount of additional knowledge. He did suddenly acquire tremendous added confidence. The result was, he instantly became more effective and more productive. When he saw himself differently, he started acting differently. He started expecting, and getting different results. Ah yes, as a man thinketh…

Zig Ziglar, See You at the Top

7 things AI can do now

/

Attitude

/

The longer I live, the more I realize the impact of attitude on life. Attitude, to me, is more important than facts. It is more important than the past, than education, than money, than circumstances, than failures, than successes, than what other people think or say or do. It is more important than appearance, giftedness or skill. It will make or break a company . . . a church . . . a home.

The remarkable thing is we have a choice every day regarding the attitude we will embrace for that day. We cannot change our past . . . we cannot change the fact that people will act in a certain way. We cannot change the inevitable. The only thing we can do is play on the one string we have, and that is our attitude . . . I am convinced that life is 10% what happens to me and 90% how I react to it. And so it is with you...we are in charge of our attitudes.

Charles R. Swindoll