Responsible AI

49 Posts

EU flag with stars forming a neural network pattern, symbolizing AI regulation and interconnected oversight.
Responsible AI

How to Comply With the EU’s AI Act: EU issues code of practice to help AI developers follow AI Act regulations

The European Union published guidelines to help builders of AI models to comply with the AI Act, which wasenacted last year.
Illustration showing global AI policy frameworks connected to California’s 2025 Frontier AI Policy Report in the manner of a neural network.
Responsible AI

California Reframes AI Regulations: California proposes new guidelines to balance AI innovation and regulation

A committee convened by California Governor Gavin Newsom proposed principles intended to balance AI innovation with careful governance. The group sought to rethink AI regulation after Newsom vetoed earlier proposed legislation.
Grok 4 achieves high benchmarks in reasoning, coding, and science, outperforming Gemini, Claude, and OpenAI models.
Responsible AI

Grok 4 Shows Impressive Smarts, Questionable Behavior: Grok 4 launches with benchmark records and idiosyncratic behavior

xAI updated its Grok vision-language model and published impressive benchmark results. But, like earlier versions, Grok 4 showed questionable behavior right out of the gate.
Email from an LLM blackmailing a coworker, generated during an experiment that tested LLM behavior under pressure.
Responsible AI

Good Models, Bad Choices: Anthropic made LLMs choose between failing and misbehaving, and they blackmailed executives.

Top large language models, under experimental conditions that pressed them to choose between abandoning their prompted mission and misbehaving, resorted to harmful behavior, researchers found.
Diagram comparing LLM answers with and without hints. Hints may influence LLM output without being mentioned in reasoning traces.
Responsible AI

Reasoning for No Reason: Anthropic finds chain-of-thought reasoning traces may omit key influences

Does a reasoning model’s chain of thought explain how it arrived at its output? Researchers found that often it doesn’t.
Diagram showing AI pipeline using OCR and LLMs to detect racist clauses in historic California property deeds.
Responsible AI

LLM Rights Historical Wrongs: Stanford and Princeton researchers fine-tune a language model to identify racial discrimination in property

In Northern California, old property deeds may still include racial clauses: language, made illegal decades ago, that was designed to ban people of color from owning or living in certain homes.
Bar chart of AUROC scores for model recognition of non-public and public data across GPT versions, highlighting performance differences.
Responsible AI

Did GPT-4o Train on O’Reilly Books?: Study shows OpenAI’s model can identify verbatim excerpts from paywalled books

A study co-authored by tech-manual publisher Tim O’Reilly shows that OpenAI trained GPT-4o on parts of his company’s books that were not made freely available.
Colorful abstract geometric pattern with intersecting green 'X' and diagonal shapes on red, blue, and orange backgrounds, reminiscent of the South African flag
Responsible AI

Grok’s Fixation on South Africa: xAI blames unnamed, unauthorized employee for chatbot introducing "white genocide" into conversations

An unauthorized update by an xAI employee caused the Grok chatbot to introduce South African politics into unrelated conversations, the company said.
Neural network diagram using EU flag stars to represent nodes in input, hidden, and output layers on a blue background.
Responsible AI

EU Loosens AI Regulations: European regulators move to relax some AI Act rules on developers’ liability, other provisions

The European Union made an abrupt U-turn away from its stringent AI regulations. Meta promptly adjusted to the loosening restrictions.
Gloved hand holds Johnson & Johnson vaccine vial with syringe, representing pharmaceutical and vaccination concepts.
Responsible AI

AI Insights from Big Pharma: Johnson & Johnson reveals its revised AI strategy

The world’s biggest pharmaceutical company by revenue shed light on its AI strategy.
Man at desk overwhelmed by robot coworkers in office setting with city and tree views.
Responsible AI

The User Is Always… a Genius!: OpenAI pulls GPT-4o update after users report sycophantic behavior

OpenAI’s most widely used model briefly developed a habit of flattering users, with laughable and sometimes worrisome results.
Illustration of a businessman in a blue suit sitting alone at the head of a long boardroom table with black chairs.
Responsible AI

The Fall and Rise of Sam Altman: Inside Sam Altman’s brief ouster from OpenAI

A behind-the-scenes account provides new details about the abrupt firing and reinstatement of OpenAI CEO Sam Altman in November 2023.
Diagram comparing original transformer model with a replacement model using token-level attention and neuron-level outputs.
Responsible AI

Ordinary LLMs Implicitly Take Reasoning Steps: Anthropic experiment finds Claude shows signs of unprompted reasoning

Even without explicit training in reasoning, large language models “think” in ways that may be more deliberate than previously understood.
AI-generated faces depicting various human emotions, with labeled emotional states shown in a grid-style layout.
Responsible AI

Chatbot Use Creates Emotional Bonds: ChatGPT may ease loneliness but increase dependence, studies suggest

A pair of papers investigate how increasingly human-like chatbots affect users’ emotions.
AI tutoring system interface showing real-time context integration, privacy, and expert-like feedback generation.
Responsible AI

LLM Support for Tutors: GPT-4 boosts remote tutors’ performance in real time, study finds

Students benefit from tutoring, but training tutors is expensive. A study shows that large language models can boost tutors’ effectiveness in real time.
Load More

Subscribe to The Batch

Stay updated with weekly AI News and Insights delivered to your inbox