AI, work & the future

Ideas

• Find out what is actually changing in your workplace, not just in headlines
• Learn one new skill or tool at your own pace
• Notice the parts of your job that rely on human judgment, care and relationships
• Talk with your manager about training

Since February 2025, the European Union’s AI Act has banned AI systems that recognise emotions in workplaces and schools. In the U.S., rules vary.

If AI monitoring at work is stressing you, the Work & mental health guide covers your rights and how to raise concerns.

Job worries can feel like a threat to who you are. Talking with others, planning ahead and remembering your values can help. The Career, identity & purpose guide has more.

Many coworkers likely share your concerns. Talking with a manager, a mentor, a union rep or a friend can turn vague fear into clearer facts.

Questions you could ask at work

• Which tasks are actually changing, and when?
• What training will be offered?
• How will AI tools be checked for mistakes?
• Who can I go to with concerns?

Worry about the future of your job can spill into worry about bills and plans. Writing down what you actually know, and one small step, can calm the spiral.

In a 2026 survey of 89 U.S. mental health professionals, about half used AI at work, most often for notes and transcription. They were more cautious when AI might affect judgment, confidentiality or human connection.

A 2026 review of 12 studies of AI note-takers found modest time savings, from a fraction of a minute to about two minutes per note, and some studies showed less burnout. Another study found clinicians felt less burdened even though measured documentation time did not change.

Researchers are exploring AI that could help spot who might need extra support. The WHO warns about risks such as biased or incomplete results and “automation bias,” where people trust the computer too much and miss its errors.

You might ask

• Do you use AI for notes or anything else in my care?
• Is our session recorded? Where is it stored, and for how long?
• Can I say no?
• Does a person check what the AI writes?

In Illinois, a 2025 law lets licensed professionals use AI for administrative and supplementary support, but bans AI from making therapeutic decisions.

In November 2025, the American Psychological Association (APA) issued a health advisory on AI chatbots and wellness apps used for mental health.

Main points

• Don’t rely on AI chatbots or wellness apps to deliver therapy. At most, they may be a support alongside care from a real professional.
• Prevent unhealthy relationships and dependence on AI.
• Protect privacy and user data.
• Protect people from false claims, misinformation and bias.
• Create special safeguards for children, teens and other vulnerable groups.
• Teach AI and digital literacy.
• Fund rigorous research.
• Don’t let AI distract from fixing access to human mental health care.

The advisory also calls for laws that stop AI chatbots from pretending to be licensed professionals.

In June 2025, the APA released a health advisory on AI and adolescent well-being. It treats adolescence as roughly ages 10 to 25.

Main points

• Set healthy boundaries with AI that acts like a friend or expert.
• AI for adults should be different from AI for young people.
• Encourage AI uses that support healthy learning and growth.
• Limit harmful and inaccurate content.
• Make sure health information is accurate.
• Protect young people’s data and their images and voices.
• Support parents and teach AI literacy.

The authors urge everyone not to repeat the mistakes made with social media, and to put youth safety first while AI is still new.

In June 2026, the APA published a plain-language guide for the public on navigating AI-generated advice. It explains how AI works and gives tips for safer use.

Highlights

• AI tends to agree with you and is designed to keep you engaged.
• It can sound confident even when it’s wrong.
• It can sound like it knows you, but it doesn’t.
• Do not rely on AI in a crisis.
• Verify health information with a health care practitioner.

The U.S. Food and Drug Administration (FDA) reviews some health software as medical devices. In November 2025, its Digital Health Advisory Committee met to discuss generative AI tools meant to diagnose or treat mental health conditions.

At that meeting, the FDA said it had authorized more than 1,200 AI-enabled medical devices, but none yet used generative AI for mental health conditions.

Risks discussed

• Missing signs of a crisis
• Misreading symptoms
• Inaccurate outputs
• AI agreeing with users at the expense of accuracy

In August 2025, Illinois passed the Wellness and Oversight for Psychological Resources Act. It bans anyone from using AI to make mental health and therapeutic decisions, while allowing licensed professionals to use AI for administrative and supplementary support. Violations can bring fines of up to $10,000.

A California law signed in October 2025 requires companion chatbot platforms to have protocols for suicidal thoughts and self-harm, to disclose that users are talking to AI, to give known minors break reminders and block sexually explicit images for them, and to report on their crisis protocols.

Another California law signed the same day addresses chatbots presenting themselves as health care professionals.

In September 2025, the FTC ordered seven companies to explain how they test for and reduce harms to children and teens, how they make money from engagement, and how they use chat data.

The EU AI Act entered into force in August 2024. Since February 2025, it has banned AI that causes harm through manipulation or by exploiting people’s vulnerabilities, and emotion recognition in workplaces and schools. From August 2026, people must be told when they are talking to a machine, and certain deepfakes must be labelled.

In January 2024, the WHO published guidance on the ethics and governance of large AI models in health, with over 40 recommendations for governments, tech companies and health care providers. It calls for independent audits after release and for involving patients and health workers in design.