More and more APIs are beginning to limit their access. For example, earlier this summer, Salesforce restricted access to the Slack API through its platform in an effort to stop organizations from using Slack data to train large language models (LLMs). Instead, users have to rely on Salesforce’s Real-Time Search API, which allows users to search Slack files and messages directly.

According to a spokesperson from Salesforce, data security was the reason for the change. “A cornerstone of Slack’s new data connectivity strategy is enabling real-time search access via our Real-time Search API. This allows users to interact with data directly where it resides, without the need to duplicate or move data and permissions between systems. This API also eliminates the need for large data exports from Slack, keeping customer data secure, while maintaining support for key use cases like permission-based search.”

More cynical users suspect that Salesforce’s transition has more to do with their hope to release their own AI tools. Either way, it’s a sign of a fragmenting API landscape and the re-emergence of proprietary solutions and data silos if left unchecked.

AI Drives Clamp Downs on Data Sharing

Salesforce’s action is just the latest example of APIs clamping down on sharing data. As Bob Hutchins of Human Voice Media recently told Computerworld Magazine, “This move by Slack/Salesforce is part of a broader pattern; we’re seeing platforms tightening their grip on user data under the banner of security or product integrity, but often in ways that primarily serve their own AI ambitions.”

He goes on to talk about some of the potential risks this trend poses. “Let’s call it what it is — platform enclosure. For organizations trying to reduce data silos, this creates friction. Many are turning to AI to surface insights across platforms — email, docs, project tools, and messaging apps like Slack. Blocking third-party access could mean fewer choices, more workarounds, and slower decision-making. If this tactic spreads to other vendors, we’re talking about a future where your data becomes less yours and more theirs.”

The rise of AI is driving many of these restrictions. For instance, Cloudflare, the internet infrastructure and security provider, recently introduced a pay-per-crawl feature, allowing website owners and content creators to opt out of machine learning and training LLMs.

Even the LLMs themselves are restricting access for other AI. Anthropic recently restricted OpenAI’s API access after discovering Claude had been used to help create GPT-5, which was a direct violation of Anthropic’s Terms of Service. Google Gemini announced a similar limitation for its API, as well as specifying how the Gemini API can be used.

These API restrictions are the latest example of APIs either severely limiting or outright restricting access to their APIs. We started seeing the API restriction trend in earnest when social media apps like Instagram, X (formerly Twitter), and Facebook limited access to their APIs. This was followed shortly by Reddit virtually eliminating its free API tier, removing many third-party applications’ ability to interact with the site.

Spotify has recently gone down a similar road, restricting API access to registered companies available in “key Spotify markets” with 250,000 or more registered users. Clearly, there’s a growing trend of digital public companies limiting access to their APIs, largely driven by AI and the drive towards API monetization.

What Happens When the Data Shuts Off?

If left unchecked, this trend towards API restriction threatens to undo much of APIs’ usefulness. If more platforms follow Salesforce’s example, it could result in data becoming siloed again. Not only will this restrict the data’s usefulness, but it will also make data harder to find, increasing the likelihood of unnecessary work and decreased developer morale.

API restrictions and data siloing threaten to stall AI agents, as well. Without access to historical and real-time external data, AI could break down into glorified bots. It also limits AI’s ability to be predictive. Without a contingency plan, API restrictions could stop agentic AI in its tracks before it fully gets a chance to begin.

Finally, siloed data brought about by API restrictions threatens many third-party apps that rely on their data. Salesforce’s Slack restrictions are already impacting Glean, a third-party application that lets users search, filter, and sort data from across their entire enterprise, significantly hampering the app’s usefulness. Some developers worry that API restrictions could result in all third-party apps being forced to sell their product exclusively through Slack’s marketplace rather than on the open market. This would greatly limit third-party developers’ competitiveness and potential reach.

Getting Ahead of the API Restriction Trend

Anyone working with APIs in any capacity would do well to get ahead of the API restriction trend. Developers might start by communicating with their existing providers, getting a guarantee about future integration and data interoperability. Users relying on Slack and Salesforce might consider seeking alternatives as another way to prepare for future data lockdowns. Developers relying on software-as-a-service (SaaS) might start thinking about transitioning away from proprietary tools, as well, reducing the risk of their data being locked in a silo where it can’t do anyone any good.

Following these steps and guidelines will help you and your team future-proof your APIs, API-driven tools, and software. Becoming familiar with the rules and regulations around proper use of data, AI, and LLMs, as well as the current best practices around data privacy and sovereignty, will help you, your team, and your entire organization be ready for anything, no matter what the future might bring.