A new report from the Adecco Group reveals that while workers believe they are saving an average of two hours a day due to artificial intelligence, this perception contrasts with the reality faced by many organizations. The study surveyed 37,500 workers across 31 countries and found that confidence in AI usage has surged, with 71 percent of respondents indicating that nothing holds them back from using AI, a significant increase from 19 percent in 2024. Despite the optimism surrounding AI, there are notable concerns; 23 percent of workers fear job loss, while a majority anticipate new job creation and changes to their roles. Denis Machuel, CEO of the Adecco Group, emphasizes the need for organizations to invest in training and ethical frameworks to ensure employees can engage confidently with AI. He states that AI adoption is now a strategic imperative, and collaboration between employers and employees will be crucial for building a resilient workforce.
Anthropic has launched its new artificial intelligence model, Claude Haiku 4.5, making it available for free to users of its Claude.ai platform. This model is priced at just $1 per million input tokens and $5 per million output tokens, significantly undercutting its predecessor, Sonnet 4, which is priced at three times more while operating faster. The release comes amid intense competition with OpenAI, as Anthropic reported a surge in its revenue, approaching $7 billion annually and targeting up to $26 billion by 2026. According to Anthropic representatives, Haiku 4.5 can handle complex tasks in parallel, allowing enterprises to deploy multiple specialized AI agents efficiently. The company emphasized that Haiku 4.5 has undergone extensive safety testing, aiming to address concerns surrounding artificial intelligence regulation and safety.
Researchers at Microsoft have made a breakthrough in battery technology, identifying a new material that could significantly reduce the lithium needed in rechargeable batteries. This discovery came from screening over 32 million potential materials and resulted in a promising candidate in just 80 hours, thanks to artificial intelligence. The Pacific Northwest National Laboratory plans to synthesize and test this novel material, known as NaxLi3−xYCl6, as part of efforts to improve battery performance and safety. Microsoft’s project leader, Nathan Baker, emphasized that the goal was to showcase how AI can accelerate scientific discovery and streamline the search for new materials. This development aligns with ongoing global efforts to create next-generation batteries that utilize more abundant materials and enhance energy storage capacity, reducing reliance on rare and expensive elements.
Why do we care?
So apparently workers think AI’s saving them two hours a day. I’d love to see the receipts. Adecco’s survey says confidence is way up—but that doesn’t mean companies are seeing those hours show up in the bottom line. Believing you’re more productive isn’t the same as being more productive. The real issue here? Training and accountability. If nobody’s measuring AI’s output quality, it’s just digital busywork.
Then there’s Anthropic—dropping Haiku 4.5, a faster, cheaper model at a buck per million tokens. That’s three times cheaper than before. The AI price war is real, and it’s going to reshape how vendors bake AI into their tools. Cheaper models mean MSPs can experiment more—but it also means you need to pick tools for quality, not just cost.
And Microsoft just used AI to find a new battery material in 80 hours. Cool, but let’s not kid ourselves—most of us don’t have 32 million data points to feed an AI lab. The point is: domain expertise plus AI works.
Here’s what to do: start tracking actual AI results, not just user feelings. Build AI literacy with your clients, and make “measurable impact” your selling point. The hype cycle’s over—this is where the grown-up AI conversation starts.