Meta chief Mark Zuckerberg wants the tech giant to go "AI-native", a term that raises questions in an era where artificial intelligence has wedged itself into the workflows of every kind of major organisation.
According to newly leaked internal documents from Meta, which set out a few of the many ways in which the company aims to achieve Mark Zuckerberg's vision for the future, the tech giant has set targets for how much AI should be used by employees for daily coding tasks.
What exactly does it mean to be 'AI-native'?
With the growing use of artificial intelligence in tech companies across the globe, being 'AI-native' has come to mean more than simply increasing AI usage to keep up with the times.
This is because tech giants that aim to increase AI usage don't just see it as an afterthought, but as a critical part of a completely new ecosystem: one that enables data-driven decision making using a blend of advanced AI and human involvement behind the scenes.
Meta's aggressive AI strategy, however, has also come at the cost of looming layoffs en masse, with a Reuters report earlier this month noting that the tech giant is planning to cut about 20 per cent of its workforce (or approximately 15,800 jobs) in the coming months.
With about 15,800 jobs at stake, the colossal cuts would surpass the 11,000-job cut in November 2022—followed by another 10,000 in the spring of 2023—as a part of what it had called the 2022-23 'Year of Efficiency', making the upcoming ones the deepest hitting cuts in the company's history, in pursuit of the AI-native aim.
On AI use for tasks
According to the newly leaked internal documents, Meta has set goals on AI usage parameters for late last year and 2026, a Business Insider report said.
The documents also involve related data on AI usage across various organisations.
One such parameter, laid out for Meta's creation org—which is responsible for building and maintaining core creative experiences—aims to have 65 per cent of engineers write more than 75 per cent of their committed code using AI. Committed code is code that has been saved and tracked in a project.
Meta's Scalable Machine Learning org, which focuses on AI models and infrastructure, had a goal for February 2026 to have AI assist 50-80 per cent of coding, the document said. However, the company has said that this goal is not being tracked via metrics.
The document also listed several company-wide goals for Q4 2025 spanning Messenger, WhatsApp, Facebook, and other major Meta products. One such goal is to have 80 per cent of mid-level to senior-level engineers adopt AI tools such as DevMate, Metamate, and Google's Gemini. For these goals, the report cites a note that the focus here is on "tool adoption" rather than increasing the percentage of code written by AI.
"It's well-known that this is a priority and we're focused on using AI to help employees with their day-to-day work," a Meta spokesperson said in the report, adding that Meta's performance programme focuses more on the impact from AI tool usage, rather than just usage.
