This week saw important steps in using AI agents to improve accessibility. In the U.S., the Access Board shared new rules for using AI at work. They said AI tools must be tested to make sure they don’t harm people with disabilities.

Microsoft showed off new AI tools at a big meeting. They made a special joystick for Xbox that helps gamers with disabilities. They also worked with other companies to create better apps for people who can’t speak.

New York City started AI training for workers. They want all city websites and apps to be easier for people with disabilities to use. This includes adding captions to videos and making buttons bigger.

In Canada, a report found many people with disabilities still face problems online. Groups are using AI helpers to fix websites and make apps work better with screen readers.

Some problems were found too. A study showed AI hiring tools sometimes treat people with disabilities unfairly. Schools and companies are being asked to check their AI for bias.

Big tech companies like Google added AI video helpers. These tools automatically write captions and let users jump to specific parts using text. This helps students and workers find information faster.

Stores got reminders to use AI accessibility checkers on their websites. New U.S. laws require all business sites to follow special rules by 2026. AI tools can scan pages and find problems fast.

Lastly, groups warned about AI assumptions. A test showed smart speakers sometimes say hurtful things to blind users. Experts want more people with disabilities to help train AI systems.