AI Arbitrator Aims to Fix an Overloaded System

The American Arbitration Association has built an AI Arbitrator using OpenAI models to handle document-only disputes, draft decisions, and explain who should win and why, always with a human arbitrator in the loop. The goal is cheaper, faster resolution for small and mid-sized businesses that often can’t afford traditional legal help.

Courts Already Quietly Rely on AI

US courts are using AI for a range of tasks: from back-office work like classifying filings and monitoring threats, to more sensitive uses such as translations, outcome prediction, and legal analysis. Older algorithmic risk tools and newer LLM-based helpers alike raise long- standing concerns about bias and reliability.

Judges Experiment With LLMs to Interpret Language

Some judges, like 11th Circuit Judge Kevin Newsom, are testing ChatGPT and other LLMs to find the “ordinary meaning” of disputed terms in cases, treating AI outputs as one data point among dictionaries and other sources. Researchers, however, say this misunderstands how models are trained and risks importing hidden biases and non-representative language patterns into judicial reasoning.

Hallucinations and Bias Threaten Access to Justice

Studies show legal-focused LLM tools still hallucinate “widespread” false legal facts, even when wrapped in systems like RAG, and that these errors fall hardest on under-resourced or self-represented litigants who most depend on free or cheap AI help. Rather than closing the justice gap, poorly governed AI could worsen existing inequalities.

Human Judgment Remains Essential

Supporters argue that human judges are also biased and inconsistent, and that AI systems that “show their work” could reduce some human errors and free people to focus on complex cases. Critics counter that today’s LLMs still lack true understanding, transparency, and stable behavior, making them too fragile to trust as core decision-makers in matters of rights and liberty.

Text Prompts Replace One‑Size‑Fits‑All Effects

Google Photos now lets users describe in words how a still image should animate, including movement, style, or visual effects, instead of relying only on preset options like “Subtle movement” and “I’m feeling lucky.”

More Creative Guidance with Suggested Prompts

Alongside the new text box, Google Photos surfaces suggested prompts that users can tap and then customize, making it easier to experiment without having to think of ideas from scratch.

AI-Generated Clips Can Now Include Audio by Default

Videos created from photos may now ship with sound automatically, so short AI clips feel more like finished, share‑ready videos instead of silent loops that require extra editing.

Feature Is Rolling Out with Age and Regional Limits

The enhanced Photo to video tool is rolling out on Android and iOS through the Create tab, but availability varies by region, and text prompts are restricted to users 18 and older.

Part of a Bigger Push for AI Media Tools

The update builds on Google’s Veo 3 video model and broader photo‑to‑video efforts across Gemini and Flow, as the company races to make AI‑powered video creation mainstream across its consumer apps.

Start a New Chat

Switch Your AI Model

Try This Prompt & Observe the Output