OpenAI has released a new economic analysis detailing the significant impact of its AI tools, particularly ChatGPT, on productivity across various sectors. According to the report, over half a billion people globally use OpenAI’s AI tools, sending more than 2.5 billion messages daily, with 330 million of those messages coming from the United States alone. The analysis, authored by OpenAI Chief Economist Ronnie Chatterji and his team, highlights that 28 percent of employed adults in the U.S. who have used ChatGPT report utilizing it for work, a substantial increase from just 8 percent in 2023. OpenAI is collaborating with experts from Harvard University and the American Enterprise Institute to assess AI’s long-term effects on productivity and the workforce.
A recent survey by Howdy.com reveals that one in six United States workers occasionally pretends to use artificial intelligence in order to satisfy their employers. The survey indicates that about 75 percent of employers expect their employees to utilize AI in some capacity, leading to significant anxiety among workers regarding their job security and the effectiveness of AI. According to the same survey, 22 percent of employees feel pressured to adopt AI tools despite lacking confidence in using them, and one-third report that learning and applying AI takes as much time as completing their tasks without it.
Leading artificial intelligence companies are reportedly unprepared for the risks posed by their own models, according to two significant reports. The Future of Life Institute and Safer AI evaluated top AI firms on their ability to identify and mitigate potential catastrophic risks, such as bio- or cyber-terrorism. Both organizations expressed concerns that even the most prepared companies exhibit troubling gaps in their plans regarding existential risk. Notably, Anthropic received the highest scores in both reports, earning a C+ grade from the Future of Life Institute and a score of 35 out of 100 from Safer AI. The reports highlight that none of the evaluated firms have developed a coherent and actionable strategy for managing the increasing power of their systems.
Kyndryl, a global enterprise technology services provider, has launched the Kyndryl Agentic AI Framework, designed to facilitate the deployment of agentic artificial intelligence alongside human teams. This initiative comes as the adoption of generative artificial intelligence has surged, prompting significant investment in technologies like large language models and inference engines. Ismail Amla, senior vice president of Kyndryl Consult, emphasized the importance of implementing agentic AI in secure, enterprise-grade environments, stating that it can be deployed similarly to Kyndryl’s infrastructure for various industries, including banking and government. The framework aims to streamline complex data insights and enhance operational efficiency, addressing critical issues such as bias and security. Kyndryl is already collaborating with clients in sectors like government and financial services to develop tailored AI solutions that improve processes such as fraud management and customer experience.
Why do we care?
OpenAI’s out there flexing big numbers—half a billion users, 2.5 billion messages a day, 28% of U.S. workers using ChatGPT for work. Sounds impressive, right? Except… those are usage numbers. They don’t tell us if it’s actually driving productivity or just creating more noise.
And get this—a Howdy.com survey says one in six workers are faking AI use just to keep their bosses happy. That doesn’t scream “game-changing” to me.
Now, Kyndryl’s Agentic AI Framework? That’s more interesting. They’re helping enterprises actually deploy AI in secure, enterprise-grade ways. For MSPs, this is the playbook: don’t sell the hype, deliver the outcomes. And let’s not forget—two watchdog reports say AI vendors are nowhere near ready for the risks their systems create. Bottom line? There’s an opportunity to help customers navigate AI, but it’s not about who’s first to use the tool—it’s about who can make it work safely and effectively. Be the AI shepard, guiding clients to solutions while looking out for their needs.