Marc Watkins

They aren’t all dead and they aren’t mini-LLMs. It’s a persona. They tasked the AI to “respond like X” and it mimics the authorial style based on its training data. Anne Lamott was the living author it produced for me. I don’t think they bothered asking her.

The Chronicle asked me to offer my prediction for what 2026 has in store for higher education and what I hope Universities will focus on our AI era. www.chronicle.com/article/high...

I did. I’m not sure what grounds Canvas would have. The only C&D he’s posted about was related to the name. Everything about this screams scam.

Einstein & The Rise of Nuisance Tech AI may be one of the most transformative innovations of our lifetime, but not when it is treated as spam.

Einstein isn’t really a threat to academic integrity so much as it is another distracting piece of marketing hype. It isn’t clear if it was a hoax, a failed app launch, or piece of tech posing as performance theater. AI agents represent challenges to online learning open.substack.com/pub/marcwatk...

I'm working on a post about the Einstein AI agent that claimed it can do a whole course for you and log into canvas. It is likely a hoax or failed vibe-coded app. It has been taken down. Agentic AI like Perplexity's Comet browser CAN take a course for you. Knowing what is BS will always be valuable

Screenshot of an article header from a website. The title reads "Writing As Thinking—By Proxy" in bold serif font. Below it, the author's name "by Jon Ippolito" appears as a red hyperlink, followed by the date "Wednesday, February 18th 2026" in gray monospace font. The article preview shows a photo of a cream-colored t-shirt on a hanger printed with cartoon robots and the text "The Transformers: Writing Instructors in the Age of A.I." alongside an italic abstract that reads: "In this provocation, Jon Ippolito questions what human capabilities AI extends and what capabilities it removes. In doing so, he charts the evolution of human writing processes alongside technology while speculating on what future human writing practices will look like."

Will “writing as thinking” survive the AI age? A provocation from Jon Ippolito, followed by a conversation among the other "Transformers," Mark Marino, @anetv.bsky.social, @mahabali.bsky.social @marcwatkins.bsky.social Jeremy Douglass, and me.

preview.electronicbookreview.com/gatherings/t...

Before You Buy AI for Your Campus, Read This Does your institution need to buy AI?

There's much talk about universities buying expensive AI licenses, but will students even use them? Arguments about access, equity, and security are often part of the marketing used to sell AI, but these become meaningless if students don't stop using free AI open.substack.com/pub/marcwatk...

We left Oxford yesterday after five days without power from the ice storm. I think it may have been restored at our house, but over half the town is still without power. The outage map tells the tale.

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Day 8/10 sharing provocative scenarios to spark conversation about AI grading. Faculty have hidden directions to catch students cheating with AI. What happens when students do the same and use prompt injection to fool AI grading tools into giving them a higher grade open.substack.com/pub/marcwatk...

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Day 7/10 sharing provocative scenarios to spark conversation about AI grading. How will parents and students feel about AI being used to grade? Some might welcome the efficiency of it, while others might ask why they're paying for college for a machine to grade them open.substack.com/pub/marcwatk...

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Day 6/10 sharing provocative scenarios to spark conversation about AI grading. Agentic AI automates workflows. When grading and feedback are viewed as just another workflow, how easy is it for faculty to cede agency to an algorithm? open.substack.com/pub/marcwatk...

Prep for the SAT with practice tests in Gemini Gemini now offers full-length, no-cost practice SATs, featuring content from The Princeton Review.

I can see so many high school students going to Gemini to take practice tests, repeatedly. Will it help? Maybe. It also might mean we're headed to a world where people become more interested in assessing knowledge in ways that are less transactional. blog.google/products-and...

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Day 5/10 sharing provocative scenarios to spark conversation about AI grading. Some faculty are using AI to proctor oral exams, recording students, what responsibilities do faculty and institutions have if sensitive information is disclosed by a student to the AI? open.substack.com/pub/marcwatk...

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Day 4/10 sharing provocative scenarios to spark conversation about AI grading. AI models offer different levels of access and capability. Students may have access to better models, use 'councils of LLMs' to grade, or argue human judgement is biased vs. AI. open.substack.com/pub/marcwatk...

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Day 3/10 of sharing provocative scenarios to spark conversation. Students may soon arrive in college expecting faculty will use AI for feedback and grading because their teachers used AI in high school. How would this affect faculty decisions around using AI? open.substack.com/pub/marcwatk...

I went through age verification instead of switching to VPN to login to Bluesky. IMO, it is absurd some states (I'm in MS) implemented this for social media, but so it goes. Equity, access, and protecting teens pose lots of challenges for online spaces. I don't know how this will work with AI.

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Day 2/10 of sharing provocative scenarios to spark conversation. Using AI to create a virtual 'twin' is already being done by some faculty. The scenario below is for a fully online, asynchronous course. What guidelines and guardrails should be put in place? open.substack.com/pub/marcwatk...

Sure, Antonio. That's fine.

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

Does your uni have guidelines or policy about using AI to grade student work? I think they should. Over the coming days, I'm sharing 10 provocative scenarios to spark conversation and hopefully convince folks in education that we need to start taking this seriously. open.substack.com/pub/marcwatk...

What We Give Up When We Let AI Decide Automation Is Easy. Judgment Is Not.

I hope those in education consider what AI grading represents beyond narratives of efficiency. Assessing how a student learns is an act of critical judgement. I’d hate to see that outsourced to a machine. open.substack.com/pub/marcwatk...

We Have Something Important to Offer Students: Agency If higher ed becomes the transactional model, only now with AI, we’re in big, big trouble.

Lot of wisdom here from @biblioracle.bsky.social . I think many of us, myself included, felt like our undergraduate experience was one of searching and failure. Education is finding yourself and we shouldn’t give up that in the name of efficiency www.insidehighered.com/opinion/colu...

Elon Musk's Grok AI floods X with sexualized photos of women and minors When contacted by Reuters for comment by email, xAI replied with the message "Legacy Media Lies."

Happening on X currently--each time a woman posts something the embedded AI allows other users to abuse them by putting them in a bikini or removing their clothing in response. And this is a feature use case Musk is promoting! www.reuters.com/legal/litiga...

What Agency Means in the Era of Automation We’ve reached the end of 2025!

My final post of 2025: stop chasing AI updates. The skills we need are human, not technical. Focus instead on exploring what agency means personally to you and students in this age of automation. open.substack.com/pub/marcwatk...

If AI Can't Stop a Student From Cheating, How Can It Ever Be Safe? Taking stock of AI in 2025 has been exhausting.

Many blame students for using AI to cheat, but we should direct our anger toward an AI industry that markets itself as being safe and responsible while refusing to implement guardrails that stop students from using these tools to impersonate learning and committing academic fraud.

Within Gemin's Pro plan ($20/month) a user can toggle on/off the watermark identifying AI-generated content. What's the purpose of watermarking media to show that it is AI generated, if paid access allows you to treat the feature like a light switch?

Gemini for Students — get free Google AI Pro for a year Get more Gemini with Google AI Pro free for a year to unlock Gemini's power for school & beyond, from homework help to video creation.

gemini.google/students/

This is the 3rd time Google has offered a Gemini trial for college students since May. They're clearly tying trials around the academic calendar. It is no wonder college instructors are hostile toward AI companies pushing free premium AI for college students during finals.

😂 When writing about AI meets the carousel of iStock and Getty images.

Advice | Can Educators Counter ‘Agentic AI’? If academics don’t deal with these problematic tools now, we’ll lose online education.

www.chronicle.com/article/can-...

My latest for the Chronicle. Online education has its many flaws and it's clear ChatGPT, and now agentic AI, broke assessments, but there's too much at stake to give up the equity and accessibility afforded by online learning.
www-chronicle-com.umiss.idm.oclc.org/article/can-...