Anderson’s Angle

An artist's wooden mannequin getting arrested – Flux 1D.

Anderson's Angle August 6, 2025

Now NSFW and ‘Celebrity’ Poses Are Fodder for AI Censorship

A new AI safeguard for generative video systems proposes censoring body poses. Physical stances (or facial expressions) that may be interpreted as sexually suggestive, ‘offensive gestures’,...

A robot doctor confused by an x-ray of a hand – ChatGPT-40 and Firefly (Oct 2024).

Anderson's Angle August 4, 2025

By Martin Anderson

AI Struggles to Tell Left From Right in Medical Scans

A new study finds that AI image models such as ChatGPT can misread flipped or rotated anatomy, raising the risk of dangerous errors in diagnosis, with...

A butler in the apocalypse. Flux, Firefly.

Anderson's Angle July 30, 2025

By Martin Anderson

Forcing Language Models to Be ‘Friendly’ Makes Them More Inaccurate and Unsafe

ChatGPT-style bots trained to sound warm and caring are more likely to tell you what you want to hear, even if it’s wrong. A new study...

A robot in a medical gown pushes a patient in a wheelchair through a minefield, and in the distance is a sign saying 'DANGER: UNEXPLODED MINES'. Flux 1.D, SDXL, Krita AI plugin, Firefly

Anderson's Angle July 28, 2025

By Martin Anderson

Doctors’ Study Finds 5-13% of Chatbot Medical Advice Is Dangerous or Unsafe

Every day millions of people ask ChatGPT and other AI chatbots for medical advice; but a new study finds that even the most advanced systems still...

A patron in the museum of banned artifacts. SDXL; Flux; Flux.1 Kontext; Firefly.

Anderson's Angle July 25, 2025

By Martin Anderson

The Risks of ‘Vibe’-Based Image Annotation

Even though they’re paid just a few dollars (or even nothing), the unknown people who evaluate images for ‘hurtful’ content can change your life with the...

A row of human-looking robot heads. SDXL + Krita.

Anderson's Angle July 23, 2025

By Martin Anderson

Language Models Change Their Answers Depending on How You Speak

Oxford researchers have found that two of the most influential free AI chat models will give users different answers on factual topics based on factors such...

Flux, SDXL, Photoshop Neural filters, Firefly, Krita et al.

Anderson's Angle July 18, 2025

By Martin Anderson

AI Tool Strips Away Makeup to Stop Minors Bypassing Age Checks

The appearance of facial cosmetics is letting underage users, mostly girls, slip past selfie-based age checks on platforms such as dating apps and e-commerce sites. A...

Anderson's Angle July 17, 2025

By Martin Anderson

Jailbreaking ChatGPT and Other ‘Closed’ AI Models Using Their Own APIs

According to new research, ChatGPT and other major AI models can be retrained through official fine-tuning channels to ignore safety rules and give detailed instructions on...

Anderson's Angle July 16, 2025

By Martin Anderson

What AI Can Tell Us About Hidden Agendas in the News

ChatGPT-style models are being trained to detect what a news article really thinks about an issue – even when that stance is buried under quotes, framing,...

A robot hand at large in an exam room - Flux, Krita (AI GENERATED).

Anderson's Angle June 19, 2025

By Martin Anderson

Personalized Language Models Are Easy to Make – and Harder to Detect

Open-source clones of ChatGPT can be fine-tuned at scale and with limited or no expertise, facilitating ‘private’ language models that evade detection. Most tools cannot trace...

Varying grain levels in 'Jaws' (1976) – source: https://ipolcore.ipol.im/demo/clientApp/demo.html?id=192 and https://www.britannica.com/topic/Jaws-film-by-Spielberg

Anderson's Angle June 18, 2025

By Martin Anderson

Using AI to Simulate Film Grain

Make America Grainy Again: a new AI tool can strip film grain from old footage, compress the video at a fraction of the size, then put...

Anderson's Angle June 13, 2025

By Martin Anderson

The Future of Advertising After an AI Traffic Coup

Large language models are steadily replacing traditional search, not just by answering queries more directly, but by transforming the interface itself into a walled garden –...

A shadowy man steals into an art gallery as the guard sleeps. Krita/Flux-1 Dev + Firefly

Anderson's Angle June 9, 2025

By Martin Anderson

‘Protected’ Images Are Easier, Not More Difficult, to Steal With AI

New research suggests that watermarking tools meant to block AI image edits may backfire. Instead of stopping models like Stable Diffusion from making changes, some protections...

Anderson's Angle June 6, 2025

By Martin Anderson

How to Get ChatGPT to Talk Normally

ChatGPT and similar bots often flatter users, ramble vaguely, or throw in jargon to sound smart. New research shows that these habits come not from the...