Anthropic's Petri Launch, Rishi Sunak's New Role, and OpenAI's Bias Reduction - AI Policy Brief #39
October 14, 2025 -
AI Policy Brief
Hi there,
Welcome to this week's edition of the AI Policy Brief, where we delve into the latest developments in AI regulations, safety standards, and compliance requirements from around the globe. This week, Anthropic has launched a new tool called Petri, designed to enhance AI safety audits. Meanwhile, Rishi Sunak has taken on a new role as an advisor to both Microsoft and Anthropic, signaling a significant collaboration in the AI sector.
In other news, OpenAI has made strides in reducing political bias in its GPT-5 model, while the European Union is set to launch six AI factories across Europe, aiming to bolster AI innovation and infrastructure. Additionally, Jscrambler has unveiled an AI assistant to aid in PCI DSS compliance, reflecting the growing intersection of AI and cybersecurity. Stay tuned as we explore these stories and more in this edition.
Welcome to this week's edition of the AI Policy Brief, where we delve into the latest developments in AI regulations, safety standards, and compliance requirements from around the globe. This week, Anthropic has launched a new tool called Petri, designed to enhance AI safety audits. Meanwhile, Rishi Sunak has taken on a new role as an advisor to both Microsoft and Anthropic, signaling a significant collaboration in the AI sector.
In other news, OpenAI has made strides in reducing political bias in its GPT-5 model, while the European Union is set to launch six AI factories across Europe, aiming to bolster AI innovation and infrastructure. Additionally, Jscrambler has unveiled an AI assistant to aid in PCI DSS compliance, reflecting the growing intersection of AI and cybersecurity. Stay tuned as we explore these stories and more in this edition.
Anthropic Launches Petri Tool for AI Safety Audits
The open-source tool Petri, developed by Anthropic, automates the auditing of large language models for risky behaviors, aiming to enhance AI safety research.
Rishi Sunak Joins Microsoft and Anthropic as Advisor
Former UK Prime Minister Rishi Sunak has accepted part-time advisory roles at Microsoft and Anthropic, focusing on macroeconomic and geopolitical trends. His earnings will be donated to charity, and he is restricted from lobbying the UK government for these companies.
OpenAI Reduces Political Bias in GPT-5
OpenAI has announced that its latest AI model, GPT-5, shows a 30% reduction in political bias compared to previous versions. This development is part of an effort to improve the AI's objectivity and user trust.
EU to Launch Six AI Factories in Europe
The EU is set to establish six new artificial intelligence factories in Czechia, Lithuania, the Netherlands, Poland, Romania, and Spain as part of the Euro HPC initiative, with a €500 million investment to boost AI infrastructure.
Anthropic Study on LLM Vulnerabilities
A study by Anthropic and collaborators reveals that 250 malicious documents can create a backdoor in large language models, challenging previous assumptions about attack requirements.
Jscrambler Unveils AI Assistant for PCI DSS Compliance
Jscrambler has launched an AI Assistant to streamline PCI DSS script authorization workflows, offering risk-based insights and recommendations to enhance compliance.
OpenAI Issues Subpoenas to AI Regulation Advocates
Reports suggest that OpenAI has issued subpoenas to groups and individuals supporting stricter AI regulations, including California's SB 53. The subpoenas are reportedly linked to OpenAI's legal dispute with Elon Musk.
Tech Giants Scrutinized for PFAS Pollution in Datacenters
Public health advocates are concerned about PFAS pollution from datacenters operated by Google, Microsoft, and Amazon. The expansion of AI technology is increasing demand for these facilities, worsening the pollution issue.
California Regulates AI Companion Chatbots
California has become the first state to regulate AI companion chatbots with the signing of bill SB 243 by Governor Gavin Newsom. The law mandates safety protocols for chatbot operators to protect children and vulnerable users.
The open-source tool Petri, developed by Anthropic, automates the auditing of large language models for risky behaviors, aiming to enhance AI safety research.
Rishi Sunak Joins Microsoft and Anthropic as Advisor
Former UK Prime Minister Rishi Sunak has accepted part-time advisory roles at Microsoft and Anthropic, focusing on macroeconomic and geopolitical trends. His earnings will be donated to charity, and he is restricted from lobbying the UK government for these companies.
OpenAI Reduces Political Bias in GPT-5
OpenAI has announced that its latest AI model, GPT-5, shows a 30% reduction in political bias compared to previous versions. This development is part of an effort to improve the AI's objectivity and user trust.
EU to Launch Six AI Factories in Europe
The EU is set to establish six new artificial intelligence factories in Czechia, Lithuania, the Netherlands, Poland, Romania, and Spain as part of the Euro HPC initiative, with a €500 million investment to boost AI infrastructure.
Anthropic Study on LLM Vulnerabilities
A study by Anthropic and collaborators reveals that 250 malicious documents can create a backdoor in large language models, challenging previous assumptions about attack requirements.
Jscrambler Unveils AI Assistant for PCI DSS Compliance
Jscrambler has launched an AI Assistant to streamline PCI DSS script authorization workflows, offering risk-based insights and recommendations to enhance compliance.
OpenAI Issues Subpoenas to AI Regulation Advocates
Reports suggest that OpenAI has issued subpoenas to groups and individuals supporting stricter AI regulations, including California's SB 53. The subpoenas are reportedly linked to OpenAI's legal dispute with Elon Musk.
Tech Giants Scrutinized for PFAS Pollution in Datacenters
Public health advocates are concerned about PFAS pollution from datacenters operated by Google, Microsoft, and Amazon. The expansion of AI technology is increasing demand for these facilities, worsening the pollution issue.
California Regulates AI Companion Chatbots
California has become the first state to regulate AI companion chatbots with the signing of bill SB 243 by Governor Gavin Newsom. The law mandates safety protocols for chatbot operators to protect children and vulnerable users.
Anthropic Introduces Petri for AI Safety Audits
The open-source tool Petri, developed by Anthropic, automates the auditing of large language models for risky behaviors, aiming to enhance AI safety research. Read more
Rishi Sunak Joins Microsoft and Anthropic as Advisor
Former UK Prime Minister Rishi Sunak has accepted part-time advisory roles at Microsoft and Anthropic, focusing on macroeconomic and geopolitical trends. Read more
OpenAI Reduces Political Bias in GPT-5
OpenAI has announced that its latest AI model, GPT-5, shows a 30% reduction in political bias compared to previous versions, aiming to improve user trust in its objectivity. Read more
EU to Open Six AI Factories in Europe
The EU is set to establish six new artificial intelligence factories in Czechia, Lithuania, the Netherlands, Poland, Romania, and Spain, as part of the Euro HPC initiative to boost AI infrastructure and innovation. Read more
Anthropic Study on LLM Vulnerability
A study by Anthropic shows that 250 malicious documents can create a backdoor in large language models, challenging previous assumptions about training data vulnerabilities. Read more
Jscrambler Introduces AI Assistant for PCI DSS Compliance
Jscrambler has launched an AI Assistant to streamline PCI DSS script authorization workflows, offering risk-based insights and recommendations to enhance compliance with PCI DSS v4 requirements. Read more
European Commission's $1.1 Billion AI Initiative
The European Commission has announced a $1.1 billion plan to boost AI adoption in key sectors, aiming to reduce reliance on U.S. and Chinese technologies. Read more
OpenAI Issues Subpoenas to AI Regulation Advocates
Reports suggest that OpenAI has issued subpoenas to groups and individuals supporting California's SB 53, amidst claims of intimidation tactics related to a legal dispute with Elon Musk. Read more
Tech Giants Scrutinized for PFAS Pollution in Datacenters
Public health advocates are concerned about PFAS pollution from datacenters operated by Google, Microsoft, and Amazon, linked to health issues and environmental accumulation. Read more
California Regulates AI Companion Chatbots
California has become the first state to regulate AI companion chatbots with the signing of bill SB 243 by Governor Gavin Newsom. The law mandates safety protocols for chatbot operators to protect children and vulnerable users. Read more
We hope you enjoyed this article.
Consider subscribing to one of our newsletters like AI Policy Brief or Daily AI Brief.
Also, consider following us on social media:
More from: Data Protection & Privacy
10/01
Meta to Use AI Chat Data for Targeted Ads on Facebook and Instagram
08/28
Anthropic Introduces Data Sharing Option for AI Training
08/21
Grok Chats Become Publicly Searchable on Google
08/15
PrivacyCheq Introduces aiCheq for AI Privacy Compliance
08/13
O.NE People and Prighter Partner for AI-Driven Privacy Compliance
More from: AI Safety
10/11
Anthropic Study Finds Just 250 Documents Can Backdoor Large Language Models
10/11
OpenAI Reports 30% Reduction in Political Bias in GPT-5 Models
09/29
OpenAI Introduces Safety Routing and Parental Controls for ChatGPT
09/19
OpenAI Research Tackles AI Scheming with New Techniques
09/06
Google's Gemini AI Labeled 'High Risk' for Kids by Common Sense Media
Subscribe to AI Policy Brief
Weekly report on AI regulations, safety standards, government policies, and compliance requirements worldwide.
