When LLM agents talk to each other, a new failure emerges: agents abandon their assigned roles and mimic their conversational partner instead. Rates reach as high as 70% with major models — and increased reasoning doesn’t help.
🧵 arxiv.org/abs/2511.09710
Our work on "Echoing" was accepted at the Agents in the Wild workshop at ICLR 2026
When LLM agents interact without oversight, they can abandon their assigned roles — and standard metrics won't catch it.
#ICLR2026 #FutureOfAI #EnterpriseAI
Poisoning the Well: Search Agents Get Tricked by Maliciously Hosted Content sforce.co/40Z1Gi5
AI search agents can be fooled by planted fake content, ~80% of queries returned the attacker's answer in controlled tests.
Authors: Shafiq Joty, Xuan Phi Nguyen, Shrey Pandit, Yifei Ming
(🧵 1/5) MAS-ProVe — the first systematic study of process verification for multi-agent systems.
Paper: arxiv.org/abs/2602.03053
Do multi-agent systems actually benefit from automatic process-level verification? This paper puts that question to a rigorous test.
(🧵 1/5) InterviewSim: A Scalable Framework for Interview-Grounded Personality Simulation bit.ly/3P4cQj9
🔍 Can LLMs faithfully simulate real people? Most evaluations never check against what individuals actually said.
@salesforce.com #EthicalUse team is hiring a Dir./Sr. Dir. of #ResponsibleAI!
The ideal candidate has:
- PM Leadership: Proven industry experience with AI/ML platforms, GAI, agents
- RAI Expertise: Deep familiarity with #RAI concepts, governance, & assurance
careers.salesforce.com/en/jobs/jr32...
New Article Alert By me @fastcompany.com 🚨
"Digital surveys may have hit the #AI point of no return."
Quantitative surveys can now be taken by AI that undetectably mimics humans. This can compromise results that will impact decisions based on these outcomes. A way forward? Qualitative Research!
1/ Interesting @profgalloway.com conversation with Tristan Harris on AI risks. From companion bots over job losses to AGI. I don’t agree with all of Tristan Harris’ views, but he excels at painting a very clear and simple picture of the risks we face.
www.youtube.com/watch?v=MLvx...
Washington’s AI Task Force has issued interim recommendations including NIST-aligned principles, dataset transparency rules, high-risk governance requirements, and new AI reporting standards. Source: www.atg.wa.gov/news/news-re... #AIPolicy
A Munich court ruled that OpenAI violated German copyright law in a case brought by GEMA over training on protected song lyrics, ordering the company to pay damages. mashable.com/article/open... #AIPolicy
what really gets me every damn time is why "AI" products built with children in mind are rarely vetted or regulated. even simple household appliances go through so much testing, assurance, & regulatory process
and now the industry is lobbying so hard, we've regressed back to a decade
Proud to be a contributor to @mtlaiethics.bsky.social's State of #AI #Ethics Report Vol. 7! 58 global contributors, 17 topics, & 48 essays. It features grounded examples, community-driven solutions, & insights from across sectors. 🌍
#responsibleAI #AIgovernance #agenticAI
👉Read: montrealethics.ai
🎉Today, we released a series of five Responsible AI courses that you can take for free: alltechishuman.org/rai-courses
Principles of Responsible AI
The History of Responsible AI
Operationalizing AI Governance (2)
Governing Agentic AI Systems
Our latest Responsible Tech Guide is out! We cover the state of Responsible Tech, pathways and pipelines, and the intersections between Responsible AI, Trust & Safety, and Public Interest Technology.
DOWNLOAD: alltechishuman.org/responsible-...
I'm excited about #NVIDIA #GTC DC this week! Please join me, Elham Tabassi (@brookings.edu), Chris Meserole (Frontier Model Forum), and @roar817.bsky.social (NVIDIA) for our panel on "Foundational Concepts in #AISafety on Wed., Oct. 29th at 3p ET!
www.nvidia.com/gtc/dc/sessi...
#AI #ResponsibleAI
32 people standing and sitting in a ballroom in front of a screen with the World Economic Forum logo and the title "Faith in Action Convening: Embedding Dignity in the Intelligent Age."
At the @weforum.org Global Futures Council Summit last week, I was honored to speak at the "Faith in Action: Embedding Dignity in the Intelligent Age" dinner. The event convened global leaders working at the intersection of #AI & moral insight to explore ways to embed human dignity in AI development
I'm thrilled to announce that @salesforce.com has obtained the ISO/IEC 42001 certification for our #AI Management System (AIMS), the first internationally recognized standard for #ResponsibleAI governance! 🎉 www.salesforce.com/blog/salesfo...
Pivot is going on tour!
Catch us in: Toronto, Boston, NYC, DC, Chicago, San Francisco and LA.
Kara Swisher. Scott Galloway. 7 Cities. 7 Days. Endless jokes and bickering.
Go to PivotTour.com for tickets!
Big news! I've been working with @sladner.bsky.social on "Off the Path" -- a limited series podcast where we talk about user research careers, especially our own. Episodes come out October 10. In this hype video, I reflect on why I decided to take on this new project together.
Excited to attend the Credo AI #AI #Trust Summit tomorrow (Tuesday)! Who else will be there?
www.credo.ai/apply-summit...
That was a fantastic event and kudos to the entire @axios.com reporting staff for asking the tough questions!
I'm excited to dial into the live stream of the @axios.com #AI + DC Summit with @inafried.bsky.social at 11a PT / 2p ET!
www.axios.com/2025/09/10/w...
A German teen told Meta's child safety researchers his under-10 little brother had been sexually propositioned multiple times on its VR platform.
Meta deleted the evidence.
New internal whistleblower docs indicate that was part of a broader cover-up: www.washingtonpost.com/investigatio...
@weforum.org GFC Report "Synthetic Data: The New Data Frontier." Bias, scarcity, & privacy constraints limit the potential of real-world data. Synthetic data offers a path forward. Learn more! @digitalarun.bsky.social @sarahooker.bsky.social @nighatdad.bsky.social
reports.weforum.org/docs/WEF_Syn...
AI companies should change the way their product works in a fundamental way. It should engage like a computer not a human, they don’t have minds, can’t think. They should work and sound like a computer. Prevent tragedy like this.
Personas & long-term memory increase the risk of people connecting emotionally with #AI & believing it is conscious. Those features may come with benefits but they also pose unintended risk to ppl's wellbeing. The "things" we risk breaking when we move fast, are PEOPLE.
www.ibm.com/think/news/h...
I’m paying you a million dollars per year. Just solve ASI already. What’s the problem?
Amen to this. I wrote about this problem in "Artificial Unintelligence" in 2018; it's only gotten worse in the AI era.
