Microsoft is taking steps to restrict its employees' use of AI.
As the use of AI in business operations becomes more widespread, a culture is emerging in some companies where employees who use AI extensively are rewarded. Amidst this trend, it has been revealed that Microsoft sent an email to its employees warning that 'simply using a lot of AI isn't always the best approach.'
Microsoft Tells Engineers 'Tokenmaxxing Is Not What We Are Optimizing For'
Microsoft tells engineers to curb their token-burning enthusiasm
https://www.theregister.com/ai-and-ml/2026/08/05/microsoft-tells-engineers-to-curb-their-token-burning-enthusiasm/5283482
'Tokenmaxxing is not what we are optimizing for': Microsoft tells engineer to calm down on AI usage | TechRadar
https://www.techradar.com/pro/tokenmaxxing-is-not-what-we-are-optimizing-for-microsoft-tells-engineer-to-calm-down-on-ai-usage
Companies that actively utilize AI are developing a culture called 'token maxing,' where employees use AI tokens as much as possible, and their usage is considered proof of productivity.
Microsoft is taking steps to curb this culture. According to internal emails obtained by overseas media outlet 404 Media, Microsoft Executive Vice President Jay Parik has notified employees to avoid consuming excessive amounts of AI tokens.
Parik stated, 'Token maxing is not what we are aiming for. I want all employees to focus on maximizing meaningful results for our customers and our business.' He continued, 'Therefore, we will update our internal guidelines and manage token spending with the same discipline as other important resources,' suggesting a change from the situation where tokens could be used freely.
Palik also stated that Microsoft will be switching to the lower-cost OpenAI GPT-5.6 AI model for internal use.
In addition, through updated internal guidelines, it was reiterated that each department at Microsoft will have an 'AI token budget target' set for July 2026, and that employees will be able to track their own AI usage.
Token maxing is also a problem for other major technology companies. Uber is a particularly well-known example, where the rapid adoption of agent-based tools like Anthropic's Claude Code and Cursor reportedly depleted its annual AI coding token budget, which had been set aside for 2026, in just four months.
Uber has imposed a limit of 'up to 240,000 yen per person per month for each AI tool' due to excessive use by its employees - GIGAZINE
Related Posts:
in AI, Posted by log1p_kr