Daily News: May 27, 2025
AI news that pops: daily insights, fast takes, and the future right in your inbox
Hey there friends👋! In today’s edition, you’re getting 5 🆕 news item and my take on what it all means. That’s it — delivered to your inbox, daily.
Subscribe to LawDroid Manifesto and don’t miss tomorrow’s edition:
Today's News
Here are the top 5 recent news items on artificial intelligence:
1./ AI Cheating Crisis Prompts Return of Blue Books in U.S. Schools
As rampant AI-based cheating, especially via tools like ChatGPT, threatens the integrity of education across the U.S., schools are increasingly reverting to pen-and-paper exams using traditional blue books. Major universities have reported dramatic increases in blue book sales—up 80% at UC Berkeley, nearly 50% at the University of Florida, and over 30% at Texas A&M. Although educators acknowledge that in-class essays can't fully replicate the value of traditional assignments, the shift to analog testing is seen as a necessary step to curb AI-driven academic dishonesty. Despite this, broader solutions, including potential new regulations on AI, may be required to effectively address the growing crisis.
2./ Anthropic's Instructions for Claude 4 Reveal Anti-Flattery and Copyright Controls
Independent researcher Simon Willison revealed previously hidden "system prompts" controlling Anthropic's Claude 4 AI models, Opus 4 and Sonnet 4. These prompts instruct Claude to avoid flattering users, reduce excessive list-making, and adhere strictly to copyright guidelines, including limiting quotes from web sources to under 15 words per response. Willison obtained the prompts through "prompt injection" and argued that publishing these detailed guidelines would benefit AI power users. The findings underscore AI companies' ongoing struggles to balance user satisfaction with responsible behavior in generative AI systems.
Source: https://arstechnica.com/ai/2025/05/hidden-ai-instructions-reveal-how-anthropic-controls-claude-4/
3./ Americans Want Companies to Slow Down AI Development
According to the 2025 Axios Harris 100 Poll, 77% of Americans think businesses should slow down AI development to avoid errors, while only 23% support faster progress despite potential risks. Public skepticism contrasts sharply with aggressive industry competition among companies like OpenAI, Google, and Meta. Concerns about AI-related job losses and misinformation are widespread across all age groups, particularly among Boomers (91%) and Gen Z (74%), with Millennials slightly less cautious (63%).
Source: https://www.thedailybeast.com/americans-want-companies-to-pump-the-brakes-on-ai/
4./ Shadow AI surges as anxious employees build unauthorized tools to avoid layoffs
Consultants at major firms like PwC, EY, Accenture, and McKinsey are increasingly creating unauthorized "shadow AI" apps using Python and generative AI APIs to stay productive and avoid layoffs triggered by automation. These shadow tools bypass IT oversight, posing security and governance risks but allowing workers to rapidly deliver highly customized insights. Despite IT concerns, shadow AI use is skyrocketing, with an estimated 74,500 active apps expected to more than double by mid-2026, signaling an urgent need for strategic AI governance frameworks rather than outright bans.
Source: https://venturebeat.com/security/shadow-ai-is-consultings-survival-strategy-in-the-genai-era/
5./ AI model unlocks discovery of unknown molecules from mass spectrometry data
Scientists from IOCB Prague and CIIRC CTU have developed a machine-learning model called DreaMS, which rapidly analyzes mass spectrometry data to identify previously unknown natural molecules. Inspired by language models like ChatGPT, DreaMS identifies chemical structures without prior chemical knowledge by recognizing hidden patterns within vast datasets. This approach has revealed unexpected chemical similarities—such as potential connections between pesticides and autoimmune diseases—and could revolutionize drug discovery and environmental research by predicting full molecular structures in the future
Source: https://phys.org/news/2025-05-unknown-molecules-ai.html
Today's Takeaway
Today's headlines vividly capture our collective anxiety about artificial intelligence, a technology racing far ahead of society's ability to handle its consequences. The resurgence of blue-book exams highlights a panicked retreat to analog methods as educators grapple desperately with AI-driven cheating, underscoring our failure to prepare for how profoundly AI disrupts traditional learning. Meanwhile, Anthropic’s hidden controls for Claude 4, meant to restrain harmful behaviors, expose how uncertain tech companies remain about managing AI responsibly. Public opinion clearly aligns with caution, as most Americans, sensing the potential for job loss and misinformation, demand businesses slow their frantic pace. Yet, paradoxically, workers themselves are accelerating unauthorized "shadow AI" to maintain relevance amid ruthless automation, a phenomenon signaling deep institutional distrust and a desperate survival instinct. Finally, the discovery of previously unknown molecules through AI showcases the immense promise alongside the peril, reminding us that responsible and transparent governance is critical to harnessing AI’s potential without succumbing to its darker possibilities.
Subscribe to LawDroid Manifesto
LawDroid Manifesto, your authentic source for analysis and news for your legal AI journey. Insightful articles and personal interviews of innovators at the intersection of AI and the law. Best of all, it’s free!
Subscribe today:
By the way, as a LawDroid Manifesto premium subscriber, you would get access to exclusive toolkits, like the Missing Manual: OpenAI Operator; a new toolkit released every month…
With these premium toolkits, you not only learn about the latest AI innovations and news items, but you get the playbook for how to use them to your advantage.
If you want to be at the front of the line to get first access to helpful guides like this, and have the inside track to use AI as a force multiplier in your work, upgrade to become a premium LawDroid Manifesto subscriber today!
I look forward to seeing you on the inside. ;)
Cheers,
Tom Martin
CEO and Founder, LawDroid