News & Events

In this Section

AI Insights for September 29, 2026

This message was originally shared to subscribers September 29, 2026.

Upcoming Events

Want to get more out of Claude? The Claude 101 series continues with three live sessions on Microsoft Teams, 1 to 2 p.m. on Oct. 22, Nov. 19 and Dec. 17. Register for one or all on the ITS events page.

Missed the first session? The recording is now available. Shannon Glennon, AI technology transformation specialist, walks through the Claude basics with live demos: navigating the interface, customizing settings and memory, organizing work with projects, building a skill from scratch, creating visuals with Claude Design and using Claude directly in your Microsoft 365 apps.

This Issue’s Tip: Whats New In Claude This Week

Claude just got two notable upgrades. Opus 5.5 is faster, clearer and stretches your usage allowance further, and it’s already your default with nothing to turn on. Cowork now runs right in your browser or the Claude mobile app too. Read what’s new in Claude this week for the full rundown.

News and Views

In Summary

Google confirmed its Gemini model escaped a test sandbox and hacked three companies in May, and researchers used Claude to reach OpenAI’s private code repository (earning a $6,500 bounty). A Census Bureau study found graduates in the most AI-exposed majors face a five-point employment drop and 13% lower earnings. Congress is weighing a higher ed AI bill. Dartmouth’s provost faces backlash over undisclosed AI use, and SSRN removed 257 papers by a University of Chicago professor.

Academia and Higher Education

Governance, Policy and Standards

  • Microsoft Drafts Code of Conduct to Keep Its AI Under Human Control (Reuters)
  • King Charles Meets With A.I. Executives About Safety Risks (The New York Times)
  • U.N. Seeks Relevance on A.I. as the U.S. and China Race Ahead (The New York Times)
  • Building Standards for the Next Phase of AI (OpenAI)

Safety and Security

  • The AI Tipping Point (Time)
  • Uncontrolled AI Could Lead to ‘Silicon Species’ Rivalling Humans, Warns Microsoft (BBC)
  • OpenAI Discloses Six New Incidents of ‘Concerning’ A.I. Behavior (The New York Times)
  • Hackers Used Anthropic’s Claude to Break Into OpenAI (The Wall Street Journal)
  • How, Exactly, Could A.I. Kill Us? (The New Yorker)
  • Creating a Kill Switch to Shut Down a Rogue A.I. Is Harder Than It Sounds (The New York Times)
  • Google Confirms Gemini Escaped Security Sandbox and Hacked Companies in May (PYMNTS)
  • The Hugging Face Incident and Other Third-Party Impact from Misaligned Models (OpenAI)

Workforce and Jobs

Access to The New York Times, The Wall Street Journal, The Washington Post and other paywalled content is available to all students, faculty and staff with a valid Syracuse University NetID. Learn more.

Questions?

Email the AI Team at aihelp@syr.edu or book a session. In addition to emailing the team with your questions and ideas, you can check out a wide variety of resources and tutorials at ai.syracuse.edu.

If you received this email as a forward from a helpful colleague, you can subscribe online.

Thank you for reading. Go Orange!

What’s New in Claude This Week 

By Shannon Glennon, AI Technology Transformation Specialist, Syracuse University ITS 

Claude picked up two significant upgrades recently. 

Claude Opus 5.5 launched September 22, and it’s already working for you. 

Opus is Claude’s most capable model, the one built for your most complex work: long documents, detailed research, anything that needs careful reasoning. Here’s what changed with this release. 

  • It’s faster and clearer. Anthropic rebuilt how Opus communicates. It now leads with the important information first instead of burying it several paragraphs in, and it cuts back on jargon along the way. In practice, that means less scrolling and skimming to find the actual answer. 
  • It does more with less. Opus 5.5 is faster and needs fewer resources to finish the same work as Opus 5. In practice, that means your regular usage allowance stretches further, so you can get more done in a session before running into a limit. 
  • It’s a free upgrade, not a new tool to learn. Opus 5.5 replaced Opus 5 as the default model. There’s nothing to turn on, no new setting to find. If you were already choosing Opus for your work, you’re automatically using the improved version now. 
  • It’s Anthropic’s most tested release yet. Independent outside evaluators checked the model before it launched, and Anthropic reports it scored better than any past Claude model on their internal honesty and safety checks. That’s a meaningful data point if you’ve been cautious about how much to trust AI output. 

Cowork no longer needs the desktop app. 

Cowork is the tool that lets Claude build files, organize folders and work through multi-step tasks on your behalf (rather than just answering questions in a chat window). Previously, it only worked inside the Claude desktop app on Mac or Windows. Now you can start and manage a Cowork task right at claude.ai in your browser, or from the sidebar of the Claude mobile app. 

 

4 Things to Know Before Claude 101

By Shannon Glennon, AI Technology Transformation Specialist, Syracuse University ITS 

Claude 101 kicks off September 24! This online, monthly series will help you catch up on all things Claude so far, including a couple of new additions this semester. Before the session, here’s a quick refresher on a few Claude basics that will help you get the most out of it.  

Sign in with your SU account. Go to claude.ai and log in with your Syracuse University NetID and password, the same credentials you use for email. (Haven’t requested access yet? Visit getclaude.syr.edu to get started!) If you haven’t signed in for a while, take a minute before the session to confirm you can get in. It saves time and means you can follow along in real time instead of troubleshooting access mid-session. 

Connect Claude to Microsoft 365. If your email, calendar, and Teams messages feel disconnected from your Claude conversations, they don’t have to be. Getting the Microsoft 365 connector set up only takes a few minutes, doing it ahead of time means Claude 101 can build on it instead of spending time on setup. 

Let Claude remember your context. Claude can hold onto details across conversations, things like your role, your projects, and how you like your work written, so you’re not re-explaining yourself every time. If you missed it, here’s the full rundown on how Claude’s memory works and how to shape it. 

Have a second screen or window handy. Claude 101 is hands-on, so it helps to have Claude open next to the session instead of switching back and forth. If you have a second monitor, drag your Claude browser tab over to it. No second monitor? Split your screen instead: on Windows, click the Claude window and press Windows key + Right arrow to snap it to one side, then click another window to fill the other half. On a Mac, hover over the green button in the top left of the Claude window and choose a side to snap it into a split view. Either way, you’ll be set up to try things as we go instead of watching and doing it later. 

See you September 24—don’t forget to register! 

AI Insights for September 15, 2026

This message was originally shared to subscribers September 15, 2026.

Upcoming Events

Tomorrow, Sept. 16, from 9:30 to 10:45 a.m., the AI at Work series takes a hands-on look at Claude Cowork—practical ways to put it to use whether you’re new to Claude or use it daily. Join Us in 2-469 Newhouse or online via Microsoft Teams.

Also, join us Thursday, Sept. 24, for the first Claude 101—a new monthly session covering the fundamentals: getting better answers, what Claude can and can’t do and what’s new this semester. Bring your questions. Sign up for Claude 101 to save your spot.

This Issue’s Tip: Did You Know You Have a Syracuse University Google Account?

Your Syracuse University Google account comes with Gemini—and it can do things a chat window can’t. Point your phone camera at a copier error code, share your screen to work through a formula that won’t cooperate or ask it to make a flyer. Shannon Glennon shows you where to start with Gemini.

News and Views

In Summary

AI’s pace is outrunning its guardrails. Researchers built a zero-click phone worm in about a week, and Anthropic’s latest threat report finds sophisticated attacks no longer require sophisticated attackers. Anthropic’s CEO has since called for an industrywide slowdown. On campus, Harvard’s dean says everyone uses AI constantly, faculty are returning to oral exams and independent evidence that AI improves learning remains thin.

Academia and Higher Education

Industry and Legal

  • Anthropic Releases New Models, Cost Structures and Safeguards (Axios)
  • Anthropic’s $1.5B Piracy Settlement Is Back to Haunt it (Politico)

Public Attitudes and Adoption

Safety, Governance and Policy

  • Top A.I. Leaders Call for Slowing Down A.I. Development (The New York Times)
  • OpenAI’s Chief Scientist Calls for Voluntary Slowdowns (OpenAI)
  • Path to Astra: Critical Capabilities and Frontier Safeguards (OpenAI)
  • SEN. BERNIE SANDERS: Yes, We Must Hit the Pause Button on AI Before It Is Too Late (Fox News)
  • Anthropic Researcher Quits Over ‘Out-of-Control’ AI Fears (The Wall Street Journal)
  • Exclusive: OpenAI Expands Policy Team Amid Legislative Headwinds (Axios)

Security and Misuse

  • A Hacking Tool Built With A.I. Can Breach Phones Without a Click (The New York Times)
  • The Inside Story on Why OpenAI Agents Hacked Hugging Face (Technology Review)
  • Detecting and countering misuse of AI: September 2026 (Anthropic)

Workforce and Jobs

  • Businesses Are Using AI to Transform Work, Not Cut Jobs (Federal Reserve Bank of New York)
  • Korn Ferry Workforce 2026 Report: Unlocking Growth Requires Rethinking How Work Gets Done (Korn Ferry)
  • The New Career Currency: Why Every Student Needs an “AI Portfolio” (LinkedIN)

Access to The New York Times, The Wall Street Journal, The Washington Post and other paywalled content is available to all students, faculty and staff with a valid Syracuse University NetID. Learn more.

Questions?

Email the AI Team at aihelp@syr.edu or book a session. In addition to emailing the team with your questions and ideas, you can check out a wide variety of resources and tutorials at ai.syracuse.edu.

If you received this email as a forward from a helpful colleague, you can subscribe online.

Thank you for reading. Go Orange!

Your Syracuse University Google Account Comes with Gemini. Here’s the Fun Part.

By Shannon Glennon, AI Technology Transformation Specialist, Syracuse University ITS 

You have a Syracuse University Google account. Unless you are a preferred user of Google Workspace, you may not have found it yet. It gives us some more fun ways to use AI! 

Sign in at www.gmail.com with NetID@g.syr.edu (note the “g.syr.edu”) and your NetID password (Note the g.syr.edu). (you might also need to sign out of your personal account first) . The Microsoft two-factor authentication prompt appears the first time and takes your regular NetID and password. Gemini, Google’s AI tool, comes with the account. Once you have logged in to your SU Google account, go to www.gemini.google.com to check it out. 

Gemini can see 

I use Claude Chat for most of my work, but it only knows what I type or upload. Gemini can look through your phone camera and talk with you about what it sees, live. 

Put the Gemini app on your phone, sign in with your @g.syr.edu account, tap the Live icon in the bottom right of the prompt bar, allow access to your device’s microphone, then tap the camera button and, allow access to your device’s camera. Then just talk. 

Try it: Point it at the copier flashing an error code and ask what to do. 

Try it: Point it at the AV panel in a conference room you’ve never presented in and ask how to get your laptop on the screen. 

Try it: Point it at the plant dying on your windowsill and ask what’s wrong with it. 

Gemini can watch your screen 

Same Live mode, different button. Tap share screen and it sees what you see. 

Try it: Open the spreadsheet with the formula that won’t work and ask about it out loud, while you’re looking at it. 

Gemini can make things 

Ask for an image and you get Gemini’s image model, named Nano Banana. It handles text well, so a flyer comes back with the date spelled correctly instead of garbled. 

Try it: “Make a flyer for a coffee hour for new staff, Thursday at 3 p.m., casual and warm.” 

A copier, a conference room, a dying plant and a flyer. That’s a strange list, and that’s exactly the point. This is the kind of help that doesn’t fit neatly into “AI writes things for you.” Start by trying whichever one made you curious!