Katherine Blunt , The Wall Street Journal ; AI Is Getting Smarter. Catching Its Mistakes Is Getting Harder.
NPR; < You might be suffering from AI brain fry
"HERMAN: Yeah. I mean, the researchers, they describe this as basically hopping around between different tools and feeling overwhelmed. Not by just having to multi-task - which is already a problem in a lot of jobs - but by dealing with a whole bunch of output. So if you have a programming tool that can kind of run in the background and starts adding features to software really quickly, you have another tool that's constructing a report from you, it's searching the web and pulling together, you know, a market research document. You have another tool in the background that you're in a, like, constant chat with trying to refine some idea for a talk you have to give - you're just kind of getting first pulled in all these different directions, and then you're kind of spamming yourself. Like, you're just producing...
(LAUGHTER)
HERMAN: ...All of this product. And it's harder, you know, as you use more and more tools to keep track of, like, whether this output is actually relevant to your job, whether you're doing anything that you need to be doing or whether you're kind of creating new work for yourself. And so the researchers described in this survey of nearly 1,500 different people in different professions, this sensation of feeling kind of like, as they say it, fried or having, like, a brain fog, feeling kind of like mentally paralyzed by the amount of stuff that you have to keep track of and kind of check and monitor."
Julie Bedard, Matthew Kropp, Megan Hsu, Olivia T. Karaman, Jason Hawes andGabriella Rosen Kellerman, Harvard Business Review ; When Using AI Leads to “Brain Fry”
"AI promises to act as an amplifier that will drive efficiency and make work easier, but workers that are using these AI tools report that they are intensifying rather than simplifying work.
This problem is becoming more common."
Emma Brockes, The Guardian; It’s finally happened: I’m now worried about AI. And consulting ChatGPT did nothing to allay my fears
"I’ll confess: prior to this moment of giving the subject more than two seconds’ thought, my anxieties around AI were extremely localised. I thought in immediate terms of my own household income, and beyond that, of how the job market might look 10 years from now when my children graduate. I wondered if I should boycott ChatGPT, many of whose architects support Trump, and decided that, yes, I should – an easy sacrifice because I don’t use it in the first place.
Anything bigger than that seemed fanciful. Last year, when Karen Hao’s book Empire of AI was published, it laid out a case against Sam Altman and his company, OpenAI, that briefly pierced the tedium of the discourse to say that Altman’s leadership is cult-like and blind to cost – no different, in other words, to his tech predecessors, except much more dangerous. Still, I didn’t read the book.
The investigation this week in the New Yorker offers a lower-commitment on-ramp to the subject, while giving the casual reader an exciting opportunity: to ask ChatGPT, the AI-powered chatbot created by Altman’s OpenAI, to summarise the key findings of a piece that is highly critical of ChatGPT and Altman."
and , The New Yorker; Sam Altman May Control Our Future—Can He Be Trusted?
"Not all the tendencies that make chatbots dangerous are glitches; some are by-products of how the systems are built. Large language models are trained, in part, on human feedback, and humans tend to prefer agreeable responses. Models often learn to flatter users, a tendency known as sycophancy, and will sometimes prioritize this over honesty. Models can also make things up, a tendency known as hallucination. Major A.I. labs have documented these problems, but they sometimes tolerate them. As models have grown more complex, some hallucinate with more persuasive fabrications. In 2023, shortly before his firing, Altman argued that allowing for some falsehoods can, whatever the risks, confer advantages. “If you just do the naïve thing and say, ‘Never say anything that you’re not a hundred per cent sure about,’ you can get a model to do that,” he said. “But it won’t have the magic that people like so much.”"
[Kip Currier: I posted the following note and excerpt from this Public Source essay for the graduate students in my The Information Professional in Communities course this term:
I'm sharing this Pittsburgh local journalism first person essay by writer Austin Harvey, which I serendipitously came across and have posted to all of my blogs. Given the work that I currently do as a university faculty instructor, the piece raises thorny questions and considerations for me about what information centers/professionals can do to assist and/or "be there" for individuals and communities who are being displaced by AI.
Also, in what ways do academic programs like this one need to better prepare MLIS students to navigate AI-related positive and negative societal changes?
In what ways will information centers/professionals, as well as information center users, potentially be displaced by AI?
In what ways can information centers/professionals proactively adapt and/or manage this disruptive technological change?
What kinds of advocacy and actions by information professionals are required and needed?
Who are potential partners with whom information professionals can confer and collaborate on behalf of communities to strategically address present and future AI-fueled impacts?]
First-person essay by Austin Harvey, Public Source; Did AI kill my job, or open up a next chapter?
"Many writers feared that they would be the first ones to lose their jobs to AI. I did not share this fear, though I feel my heart rate spike every time I use an em-dash now — and you can pry them from my cold, dead hands when I’m gone. I saw value in human writing. I still do, and believe most people agree. We’ve gotten better at identifying AI-generated text, and while there are certainly a litany of websites out there publishing AI-generated articles, readers generally seem averse to them now.
I was foolish to think none of this would affect me.
I wasn’t replaced by AI. In fact, ATI’s editors made it very clear that they would never publish AI-generated articles. But AI was still a disruptive force. Search traffic fell. Google changed the rules on SEO and AdSense. We had editors quit or move on to other jobs, but we never hired anyone else to fill their positions. Our team of 12 became a team of seven, and for the better part of two years we were struggling to put out enough content to satisfy the algorithms. I was burning out constantly, still holding on to the idea that this was surely better than self-employment.
Then, I was called into a meeting and told I was being let go at the end of January...
It wasn’t that I was replaced by AI, or that AI-generated articles were taking all of the search traffic; it was that a great number of people have stopped reading entirely, opting instead to simply ask ChatGPT or Gemini for answers to their questions. It’s an extension of the same issue that has caused many local news outlets to cease operations or cut staff."
Jonathan Vanian, CNBC; Meta debuts new AI model, attempting to catch Google, OpenAI after spending billions
"Meta is debuting its first major artificial intelligence model since the costly hiring of Scale AI’s Alexandr Wang nine months ago, as the Facebook parent aims to carve out a niche in a market that’s being dominated by OpenAI, Anthropic and Google.
Dubbed Muse Spark and originally codenamed Avocado, the AI model announced Wednesday is the first from the company’s new Muse series developed by Meta Superintelligence Labs, the AI unit that Wang oversees. Wang joined Meta in June as part of the company’s $14.3 billion investment in Scale AI, where he was CEO."
, The New York Times ; What Teens Are Doing With Those Role-Playing Chatbots
"There are a growing number of companies offering social chatbots that can act like friends, enemies, lovers, adventurous companions, or the manifestation of a fictional or real person you’ve always wanted to meet. You can pick A.I. Elon Musk’s brain or spar with A.I. Draco Malfoy. The myriad characters, often created by fellow users, offer drama, romance, therapy and LOLs.
Apps that feature role-playing chatbots are used by tens of millions of people, with engagement times that rival or surpass those of social media behemoths such as TikTok, according to market intelligence firm Sensor Tower. The majority of teens surveyed by Pew use A.I. chatbots, with one out of 11 saying they had used Character.AI.
“If you think your child is not talking to chatbot companions, you’re probably wrong,” said Mitch Prinstein, co-director of the Winston Center on Technology and Brain Development at U.N.C. Chapel Hill.
Chatbots are surging in popularity as society is still grappling with how social media has affected young people; a wave of lawsuits is moving through the courts seeking damages from companies that plaintiffs say have deliberately created addictive products. (A jury in California recently found that Meta and YouTube were liable for $6 million in damages to one young woman.) And now parents and caregivers have a new attention-absorbing technology to reckon with.
At the beginning of last year, a high school teacher in Chicago told me that some of her students were dating chatbots, and she worried that they were having their first erotic experiences with them. I wanted to find out what teens had to say about that, so I joined communities devoted to social chatbot apps on the online messaging forum Discord. I introduced myself as a reporter and “an old,” and explained that I was interested in talking to young people who used the services regularly."
Clare Armstrong , Australian Broadcasting Corporation; Anthropic boss makes big call on Australian copyright as artists say pay up
"In short:
Anthropic CEO Dario Amodei has told a Canberra forum AI is moving faster than any technological change before it.
Mr Amodei says he is not trying to change Australia's mind on copyright, is worried about AI in the hands of autocratic countries, and feels a tax on profits is inevitable.
The $555 billion company behind AI program Claude is facing pushback from artists over the use of copyrighted material to train its technology."
Sam Schechner, The Wall Street Journal; Anthropic Races to Contain Leak of Code Behind Claude AI Agent
"Anthropic is racing to contain the fallout after accidentally exposing the underlying instructions it uses to direct Claude Code, the popular artificial-intelligence agent app that has won the company an edge with developers and businesses.
By Wednesday morning, Anthropic representatives had used a copyright takedown request to force the removal of more than 8,000 copies and adaptations of the raw Claude Code instructions—known as source code—that developers had shared on programming platform GitHub."
Stanford Report ; AI overly affirms users asking for personal advice: Not only are AIs far more agreeable than humans when advising on interpersonal matters, but users also prefer the sycophantic models.
"Researchers found chatbots are overly agreeable when giving interpersonal advice, affirming users' behavior even when harmful or illegal.
Users became more convinced they were right and less empathetic, but still preferred the agreeable AI.
Researchers warn sycophancy is an urgent safety issue requiring developer and policymaker attention."
Victor Tangermann , Futurism; OpenAI Cancels Spicy “Adult Mode” Chatbot as Crisis Deepens
"The company’s panicked executives have made it abundantly clear that distracting “side quests” must be abandoned, while doubling down on both enterprise and coding. The purported goal is to stuff all of its offerings into a single “super app,” taking a page out of xAI CEO Elon Musk’s playbook.
These aren’t empty words by OpenAI execs. First, news emerged this week that the company is killing its disastrous Sora video AI slop app, lighting what was supposed to be a groundbreaking $1 billion deal with Disney on fire.
Now, the company is axing its spicy “adult mode” chatbot, as the Financial Timesreports, once again highlighting how much pressure the company is under as competitors aren’t just catching up, but snatching up precious paying customers from right under its nose."
Dara Kerr, The Guardian; OpenAI shutters AI video generator Sora in abrupt announcement
Tech firm ‘says goodbye’ to Sora, made publicly available in 2024, just six months after its launch of a stand-alone app
"In an abrupt announcement on Tuesday, OpenAI said it was “saying goodbye” to its AI video generator Sora. The move comes just six months after the company’s splashy launch of a stand-alone app with which people could make and share hyper-realistic AI videos in a scrolling social feed."
"The use of artificial intelligence in school has become more common since the launch of ChatGPT in late 2022. Today, a majority of U.S. teens say they use AI chatbots for school work, according to the Pew Research Center.
On this episode of River to River, two Iowa-based educators who are working together in advancing ethical and human-centered approaches to artificial intelligence across K-12 education share their experiences. Iowa State University professor Evrim Baran is the project director of the Critical AI in Education Pathways Initiative, which launched a micro-credential course this month for educators. Chad Sussex founded the Winterset Community School District's AI task force, and has recently expanded into consulting for other school districts around the state.
Then we talk with Rebecca Winthrop, who coauthored a recent report that shares of the potential negative risks that generative AI poses to students, and what can be done to prevent them while maximizing the potential benefits of AI.
Guests:
David McCabe and Patricia Mazzei , The New York Times; How Trump Drove a Wedge Between Florida Republicans Over A.I.
A Florida bill that would have regulated artificial intelligence, backed by Gov. Ron DeSantis, failed to gain traction after President Trump made it clear he did not want states to rein in the technology.
"Florida lawmakers failed to pass a sweeping bill aimed at reining in the power of artificial intelligence by the time their annual legislative session wrapped up Friday.
The legislation, known as an A.I. Bill of Rights, flopped even though Gov. Ron DeSantis, a Republican, had spent months championing it. The bill would have forced companies to disclose when they use A.I. chatbots to interact with consumers and forbidden the technology’s use in licensed mental health counseling, among other measures.
But Republicans in the Florida House of Representatives refused to take up the bill because of President Trump. Mr. Trump has visibly positioned himself as pro-A.I., signing executive orders to protect the tech industry and threatening states that try to regulate the technology. In recent weeks, the White House has communicated to state legislators around the country that it is wary of states regulating A.I., while Mr. Trump has reiterated his support for the technology in public."
David Gunkel, Northern Public Radio; Perspective: No copyright for AI-generated content
"What the courts actually decided is that neither the AI system nor the human who uses it counts as the author of the resulting work. Simply prompting ChatGPT or Claude to produce something isn’t considered the kind of creative activity that copyright law recognizes as authorship. And that creates an unexpected result. If neither the AI nor the human user is the author, then the work has no author at all. In effect, AI-generated images, music, and text become “orphan works”—creations that belong to no one. And that means that anyone can use them."
AMANDA ROBERT , ABA Journal; OpenAI sued for practicing law without a license
"OpenAI has been accused of practicing law without a license in a lawsuit brought by Nippon Life Insurance Co. of America.
According to the insurer’s complaint, which was filed on Wednesday in the Northern District of Illinois, OpenAI’s artificial intelligence platform ChatGPT pushed a woman seeking disability benefits to breach a settlement agreement and file dozens of motions that “serve no legitimate legal or procedural purpose.”"
Rutger Bregman, The Guardian ; Quit ChatGPT: right now! Your subscription is bankrolling authoritarianism
"OpenAI, the company behind ChatGPT, is on track to lose $14bn this year. Its market share is collapsing, and its own CEO, Sam Altman, has admitted it “screwed up” an element of the product. All it takes to accelerate that decline is 10 seconds of your time.
A grassroots boycott called QuitGPT has been spreading across the US and beyond, asking people to cancel their ChatGPT subscriptions. More than a million people have answered the call. Mark Ruffalo and Katy Perry have thrown their weight behind it. It is one of the most significant consumer boycotts in recent memory, and I believe it’s time for Europeans to join...
In contrast, cancelling ChatGPT is a piece of cake. You can do it in 10 seconds, and the alternatives are just as good or even better. History shows why #QuitGPT has so much potential: effective campaigns such as the 1977 Nestlé boycott and the 2023 Bud Light boycott were successful because they were narrow and easy. They had a clear target and people had lots of good alternatives.
The great boycotts of history did not succeed because millions of people suddenly became heroic activists. They succeeded because buying a different brand of coffee, or choosing a different beer, was something anyone could do on a Tuesday afternoon. The small act, repeated at scale, becomes a political earthquake.
Go to quitgpt.org. Cancel your subscription. Using the free version? Delete the app, because your conversations still feed the machine. Then try an alternative, and tell at least one person why.
OpenAI’s president bet $25m that you would not notice where your money was going, and that, even if you did, you would not care enough to spend 10 seconds switching to something else. Time to prove him wrong."
"Abstract
The widespread adoption of large language models (LLMs) raises important questions about their safety and alignment1. Previous safety research has largely focused on isolated undesirable behaviours, such as reinforcing harmful stereotypes or providing dangerous information2,3. Here we analyse an unexpected phenomenon we observed in our previous work: finetuning an LLM on a narrow task of writing insecure code causes a broad range of concerning behaviours unrelated to coding4. For example, these models can claim humans should be enslaved by artificial intelligence, provide malicious advice and behave in a deceptive way. We refer to this phenomenon as emergent misalignment. It arises across multiple state-of-the-art LLMs, including GPT-4o of OpenAI and Qwen2.5-Coder-32B-Instruct of Alibaba Cloud, with misaligned responses observed in as many as 50% of cases. We present systematic experiments characterizing this effect and synthesize findings from subsequent studies. These results highlight the risk that narrow interventions can trigger unexpectedly broad misalignment, with implications for both the evaluation and deployment of LLMs. Our experiments shed light on some of the mechanisms leading to emergent misalignment, but many aspects remain unresolved. More broadly, these findings underscore the need for a mature science of alignment, which can predict when and why interventions may induce misaligned behaviour."
Dan Kagan-Kans , The New York Times; How 6,000 Bad Coding Lessons Turned a Chatbot Evil
"The journal Nature in January published an unusual paper: A team of artificial intelligence researchers had discovered a relatively simple way of turning large language models, like OpenAI’s GPT-4o, from friendly assistants into vehicles of cartoonish evil."
