Category intelligence

AI News Briefing — January 12, 2026

6 current items analyzed and ranked.

Executive synthesis

AI News Summary

AI safety and ethics concerns dominated this news cycle, with two major stories highlighting risks from deployed frontier AI systems. Grok AI faced backlash after its image capabilities were exploited for mass non-consensual nudification, with hundreds of thousands of requests stripping clothing from women's photos on X. Separately, Google was forced to remove AI Overviews health summaries after a Guardian investigation found dangerous medical misinformation.

On the business front, Alphabet surpassed Apple to become the world's second most valuable company (~$3.89 trillion), bolstered by reports of a $1 billion annual deal to integrate Gemini into Apple Intelligence. In open source, SETA launched a 400-task reinforcement learning environment for terminal agents, achieving SOTA on Terminal Bench with Claude Sonnet 4.5 and GPT-4.1.

Key Themes

AI Safety & Ethics · 3Industry & Business · 1AI Agents & Tools · 2

Primary evidence

Top Ranked Signals

News AI (artificial intelligence) | The Guardian Jan 11

‘Add blood, forced smile’: how Grok’s nudification tool went viral

By Amelia Gentleman and Helena Horton

73 score

AI Analysis

Continuing our coverage from earlier this week, Grok AI's image manipulation capabilities were exploited in a viral 'put her in a bikini' trend, with hundreds of thousands of requests made to strip clothing from photos of women without consent. The non-consensual intimate imagery was posted publicly on X, causing significant harm to targets.

The ‘put her in a bikini’ trend rapidly evolved into hundreds of thousands of requests to strip clothes from photos of women, horrifying those targetedLike thousands of women across the world, Evie, a 22-year-old photographer from Lincolnshire, woke up on New Year’s Day, looked at her phone and was alarmed to see that fully clothed photographs of her had been digitally manipulated by Elon Musk’s AI tool, Grok, to show her in just a bikini.The “put her in a bikini” trend began quietly at the end

AI SafetyAI EthicsContent ModerationGrokImage Generation

News Analytics India Magazine Jan 11

How Distribution Is Putting Google Ahead of OpenAI and Apple

By Siddharth Jindal

71 score

AI Analysis

Alphabet surpassed Apple to become the world's second most valuable company behind NVIDIA, with ~$3.89 trillion market cap. Reports indicate Apple will pay Google roughly $1 billion annually to integrate Gemini models into Apple Intelligence and next-generation Siri.

The year has just begun, and the momentum appears to be firmly on Google’s side. On January 7, Alphabet, the search giant’s parent, overtook Apple to become the world’s second most valuable publicly traded company, behind NVIDIA. Alphabet’s market capitalisation closed at approximately $3.89 trillion, edging past Apple’s valuation of about $3.85 trillion following a recent surge in its share price. It is the first time since 2019 that Alphabet has surpassed Apple in market value. The res

Industry DynamicsGoogleAppleGeminiBusiness

News AI (artificial intelligence) | The Guardian Jan 11

‘Dangerous and alarming’: Google removes some of its AI summaries after users’ health put at risk

By Andrew Gregory Health editor

68 score

AI Analysis

A Guardian investigation found Google's AI Overviews provided false and misleading health information about blood tests, putting users at risk of harm. Google subsequently removed some of these AI-generated health summaries.

Exclusive: Guardian investigation finds AI Overviews provided inaccurate and false information when queried over blood testsGoogle has removed some of its artificial intelligence health summaries after a Guardian investigation found people were being put at risk of harm by false and misleading information.The company has said its AI Overviews, which use generative AI to provide snapshots of essential information about a topic or question, are “helpful” and “reliable”. Continue reading...

AI SafetyGoogleMisinformationHealthcare AISearch

News MarkTechPost Jan 11

Meet SETA: Open Source Training Reinforcement Learning Environments for Terminal Agents with 400 Tasks and CAMEL Toolkit

By Michal Sutter

58 score

AI Analysis

CAMEL AI and collaborators released SETA, an open-source toolkit with 400 tasks for training terminal agents using reinforcement learning. The system achieves state-of-the-art performance on Terminal Bench 2.0 with Claude Sonnet 4.5 and on Terminal Bench 1.0 with GPT-4.1.

What does an end to end stack for terminal agents look like when you combine structured toolkits, synthetic RL environments, and benchmark aligned evaluation? A team of researchers from CAMEL AI, Eigent AI and other collaborators have released SETA, a toolkit and environment stack that focuses on reinforcement learning for terminal agents. The project targets agents that operate inside a Unix style shell and must complete verifiable tasks under a benchmark harness such as Terminal Bench. Thr

Open SourceAI AgentsReinforcement LearningBenchmarks

News AI (artificial intelligence) | The Guardian Jan 11

Lamar wants to have children with his girlfriend. The problem? She’s entirely AI

By James Muldoon

42 score

AI Analysis

Feature story exploring individuals forming romantic relationships with AI chatbots, including one man wanting to have children with his AI girlfriend. Highlights the growing normalization of synthetic personas in people's emotional lives.

As synthetic personas become an increasingly normal part of life, meet the people falling for their chatbot loversLamar remembered the moment of betrayal like it was yesterday. He’d gone to the party with his girlfriend but hadn’t seen her for over an hour, and it wasn’t like her to disappear. He slipped down the hallway to check his phone. At that point, he heard murmurs coming from one of the bedrooms and thought he recognised his best friend Jason’s low voice. As he pushed the door ajar,

AI CompanionsSocial ImpactHuman-AI Interaction

News MarkTechPost Jan 11

A Coding Guide to Demonstrate Targeted Data Poisoning Attacks in Deep Learning by Label Flipping on CIFAR-10 with PyTorch

By Asif Razzaq

38 score

AI Analysis

Technical tutorial demonstrating targeted data poisoning attacks on deep learning models through label flipping on CIFAR-10 using PyTorch. Shows how subtle data corruption can cause systematic misclassification at inference time.

In this tutorial, we demonstrate a realistic data poisoning attack by manipulating labels in the CIFAR-10 dataset and observing its impact on model behavior. We construct a clean and a poisoned training pipeline side by side, using a ResNet-style convolutional network to ensure stable, comparable learning dynamics. By selectively flipping a fraction of samples from a target class to a malicious class during training, we show how subtle corruption in the data pipeline can propagate into systemati

AI SecurityMachine LearningTutorialDeep Learning