Researchers in Italy hosted the SYNERGY 2025 workshop to improve how humans and AI solve problems together. They focused on making AI systems that adjust to human needs, like handling stress during teamwork. Over 100 experts shared ideas for building AI that asks humans for help when confused.

Google unveiled two major tools for businesses. The A2A protocol lets AI helpers from different companies (like PayPal and Salesforce) share information safely. Their AI Agent Marketplace allows businesses to buy pre-made AI tools, with options from Deloitte and UiPath launching this week. Google also added AI smarts to Workspace apps like Docs for automatic writing help.

Schools and colleges gained new AI study buddies. Clarivate’s Academic AI Platform now includes helpers that guide students through research papers and help teachers find funding. A U.K. university group praised these tools for saving time but warned about checking AI facts.

Governments are betting big on AI teamwork. The U.S. Department of Energy chose 16 sites for AI data centers powered by nuclear and solar energy. Construction starts late 2025. Meanwhile, the U.K. government’s Humphrey AI began testing features like meeting summaries and law document sorting, aiming to save £45 billion yearly.

Workers still struggle with AI partners. A Stanford report found AI handles data 30% faster than humans but makes more mistakes in emotional tasks like customer service. Companies like Amazon reduced warehouse costs using robots, but workers say machines make jobs more stressful.

Creative fields saw exciting collaborations. At Milan’s Design Week, artists used AI to create marble patterns that cut material waste by 15%. The AI suggested designs, while humans picked the best ones for real-world use.

Energy companies are building smaller nuclear reactors to power hungry AI systems. Eight big tech firms (including Google) promised to triple nuclear power by 2050 to keep AI helpers running.

While progress continues, experts remind us that human oversight remains key. MIT researchers found AI-human teams often perform worse than either alone if they don’t communicate well. Training programs teaching workers to guide AI (not fight it) are spreading in the U.S. and Europe.