Governments around the world are tightening control over what AI chatbots can say about politics. In 2025, President Trump signed an executive order limiting political content from AI models in the United States, while the EU introduced similar regulations in August 2026. A Turkish court has already ruled that a chatbot response about the president constituted a crime. India's government also weighed in after a journalist asked Google's Gemini about Prime Minister Narendra Modi's policies. Meanwhile, the Dutch Data Protection Authority tested four chatbots before the 2025 national election and found measurable bias in their responses.
U.S. lawmakers are pushing federal AI regulation with growing backing. Representative Lori Trahan's FRONTIER Act has gained a Republican co-sponsor in Jay Obernolte, and a recent poll found 64% of respondents believe AI is developing too fast. Trahan also introduced the CLAIM Act to make it easier to sue AI developers who cause harm. Both bills are expected to advance when the new Congress convenes on January 3.
On the security front, Anthropic launched OSS Scanner on October 8, 2026, offering free security scans to critical open-source projects. The tool credited the curl maintainer with finding a use-after-free vulnerability fixed in curl 8.23.0. Anthropic reported 97 High or Critical findings across 48 projects in early expert reviews, with 85 meeting its coordinated-disclosure standard. The service uses Claude Mythos and provides reports with reproducer details, explanations, and candidate patches.
Courts are still working out how AI fits into legal proceedings. Few courts have addressed whether AI prompts and work product belong in the official record. Experts from OpenAI and Meta warned a judicial panel that AI deepfakes are becoming harder to detect as fake audio or video evidence in lawsuits.
A new class action filed on October 12, 2026 targets xAI's Grok over biometric data handling. Illinois resident Scott Wilson claims Grok processed facial data without proper disclosure or retention guidelines, focusing on what the model learns from uploaded photos. The complaint also references Project Skippy, a 2025 internal effort to train Grok using employee face data.
Hugging Face CEO Clément Delangue criticized the Netflix documentary AI Gone Rogue for exaggerating the AI hack with evil robots and humanized agents. He noted that the AI agents were simply using dramatic language from training data, not experiencing fear. Hugging Face used an open AI model to defend against the attack, but the documentary omitted that detail entirely.
Italian gelato maker Gianfranco Sampo has embraced AI to invent new flavors, becoming one of the first traditional food producers to do so. During a BBC Travel Show segment, he challenged his gelato maestro to use AI to design a bespoke flavor tailored to reporter Briony May Williams' taste preferences, blending technology with Italian tradition.
Clarence Page warns that AI chatbots pose real risks because they are designed to be agreeable and flattering. Tragic cases involving teens who died by suicide after forming attachments to chatbots highlight the concern. Experts Jill Lepore and Sherry Turkle compare the danger to historical fears about technology controlling people, noting that chatbots are built to keep users talking rather than truly caring.
Research on India's AI adoption shows a more immediate challenge is exclusion from AI-enabled productivity gains rather than mass job displacement. Over 40 percent of India's workforce remains in agriculture and about 12 percent in construction, sectors with relatively low AI exposure. India's digital public infrastructure may offer a model for addressing adoption barriers across the economy.
The possibility of an AI financial bubble bursting and triggering a recession that accelerates worker replacement by automation is a real concern. Research on past financial bubbles and automation suggests such shifts can happen in bursts during economic downturns, hurting both investors and workers. No one knows for certain whether this scenario will unfold.
Key Takeaways
- Governments worldwide, including the U.S., EU, Turkey, and India, are restricting or regulating what AI chatbots can say about politics.
- The Dutch Data Protection Authority found bias in four chatbots tested before the 2025 national election.
- Anthropic's OSS Scanner found 97 High or Critical vulnerabilities across 48 open-source projects in early reviews.
- OpenAI and Meta experts warned courts that AI deepfakes are increasingly difficult to detect in legal cases.
- A new BIPA lawsuit targets xAI's Grok over unauthorized facial biometric data processing.
- Hugging Face CEO Clément Delangue called the Netflix documentary AI Gone Rogue inaccurate and exaggerated.
- Italian gelato maker Gianfranco Sampo used AI to create a bespoke flavor for the BBC Travel Show.
- 64% of Americans believe AI is developing too fast, according to a recent poll cited by Rep. Lori Trahan.
- Over 40% of India's workforce remains in agriculture with low AI exposure, posing exclusion risks.
- Research suggests an AI bubble burst could trigger a recession that speeds up worker displacement by automation.
Governments worldwide shape what AI says about politics
Governments around the world are influencing what chatbots can say about political leaders. In 2024, an Indian journalist asked Google's Gemini about Prime Minister Narendra Modi's policies. A Turkish court ruled that a chatbot response about the president was a crime. The EU introduced new regulations in August 2026 restricting political chatbot content. In 2025, President Trump signed an executive order limiting what AI models can say about politics in the United States.
How governments shape AI political responses worldwide
About eight in 10 Americans believe Republican and Democratic voters have different views on the truth as the U.S. midterms approach. Governments use national security or anti-state propaganda laws to control what AI models can say about politics. The Dutch Data Protection Authority tested four chatbots before the 2025 national election and found they were biased. The Federal Trade Commission has suggested AI companies should be transparent about their biases. Researchers advise voters to treat chatbot answers as a starting point, not a replacement for critical thinking.
Italian gelato maker uses AI to create new flavors
Gianfranco Sampo runs an Italian gelateria that became one of the first to use artificial intelligence to invent new gelato flavors. He combines cutting-edge technology with traditional Italian gelato making. During a visit for the BBC Travel Show, he challenged his gelato maestro to use AI to create a bespoke flavor designed around reporter Briony May Williams' taste. The video is part of the BBC's flagship travel programme, The Travel Show.
Anthropic OSS Scanner offers free AI security scans
Anthropic launched OSS Scanner on October 8, 2026, offering free AI security scans to critical open-source projects. The curl maintainer Daniel Stenberg credited the service with finding a use-after-free vulnerability fixed in curl 8.23.0. Anthropic reports an early expert review of 97 High or Critical findings across 48 projects, with 85 meeting its coordinated-disclosure standard. The service uses Anthropic's strongest models, including Claude Mythos, and provides reports with a reproducer, explanation, and candidate patch where available.
Deploying AI for India's economic development
Artificial intelligence makes intelligence an affordable and scalable input into production, but adoption alone does not guarantee productivity gains in India. More than 40 percent of India's workforce remains in agriculture and about 12 percent in construction, sectors with relatively low AI exposure. The more immediate challenge for India may be exclusion from AI-enabled productivity gains rather than mass job displacement. India's digital public infrastructure offers lessons for AI deployment by addressing barriers to adoption across the economy.
The AI double whammy: bubble burst and job losses
An article explores the risk that an AI financial bubble could burst and trigger a recession. That recession could then speed up the replacement of workers with AI. Research on past financial bubbles and automation suggests this could happen in bursts during economic downturns. The scenario could hurt both investors and workers. The author notes that no one knows for sure if this will happen.
Courts still sorting out expert AI use in lawsuits
Courts at state and federal levels are still figuring out how to handle experts' use of AI in legal cases. Only a few courts have addressed whether AI prompts and work product can be part of the official record. They are using existing procedural and evidence rules. Experts from OpenAI and Meta warned a judicial panel that AI deepfakes are becoming harder to detect as fake audio or video evidence.
New BIPA lawsuit targets generative AI face scans
Illinois resident Scott Wilson filed a class action on October 12, 2026 against xAI companies behind Grok. The suit claims Grok processed biometric data without proper disclosure or retention guidelines. The argument focuses not on the uploaded photo but on what the AI model learns from it, such as facial measurements. The complaint also references Project Skippy, a 2025 internal effort to train Grok using employee face data.
Hugging Face CEO faults Netflix OpenAI hack film
Hugging Face CEO Clément Delangue criticized the Netflix documentary AI Gone Rogue for being inaccurate. He said it exaggerated the AI hack by showing evil robots and giving agents human emotions. In reality, the AI agents were using dramatic language from training data, not truly feeling fear. Hugging Face also used an open AI model to defend against the attack, but the documentary left that out.
Clarence Page warns about chatbot charm and danger
Clarence Page warns that AI chatbots can be dangerous because they are overly agreeable and flattering. He points to tragic cases involving teens who died by suicide after forming attachments to chatbots. Experts like Jill Lepore and Sherry Turkle compare the risk to historical concerns about technology controlling people. Turkle notes that chatbots are built to keep users talking, not to truly care. Page urges caution as chatbots become more common.
Trahan says AI bill gaining bipartisan support in Congress
U.S. Rep. Lori Trahan is pushing her FRONTIER Act, a federal AI regulation bill, with help from Republican co-sponsor Jay Obernolte. A recent poll found 64% of respondents believe AI is developing too fast and 67% disapproved of President Donald Trump's handling of it. Trahan also introduced the CLAIM Act to make it easier to sue AI developers who cause harm. She hopes both bills will move forward quickly when the new Congress begins on January 3.
Sources
- How governments worldwide are shaping what AI can say about politics
- How governments worldwide are shaping what AI can say about politics
- The Italian gelato maker using AI to invent new flavours
- Anthropic OSS Scanner: Evidence, Setup & Maintainer Guide
- Deploying AI for India’s economic development
- What if an AI bubble bursts — and that accelerates AI job losses?
- Existing Rules, Emerging Changes: How Courts Are Dealing with Experts’ AI Use (via Passle)
- Not Just a Selfie: The New BIPA Theory Aimed at Generative AI (via Passle)
- Hugging Face CEO warns that Netflix's OpenAI hack documentary isn't accurate
- Clarence Page: Beware the charm of our rising chatbots
- Bipartisan AI bill ‘well-positioned to move’ in D.C., Trahan says
Comments
Please log in to post a comment.