The question of the limits of the capabilities of Yandex's voice assistant "Alisa" has once again come into the spotlight. Against the backdrop of discussions in the Ministry of Digital Development about potentially expanding the rights of AI assistants on devices pre-installed in Russia, the company hastened to provide clear clarifications. Boris Ryutin, head of the security group for "Alisa AI" and Yandex's self-driving transport, stated directly: the assistant has no and will not have unlimited access to user data, including banking applications.
This statement is a direct response to an initiative being discussed in government circles. It concerns the possible inclusion of "Alisa" in the list of software mandatory for pre-installation and, more importantly, granting it expanded rights on smartphones, tablets, and smart speakers. Market analysts have already expressed concern that such a step could lead to the actual monopolization of the voice assistant market in favor of Yandex.
Yandex's position: "everything is under control"
Ryutin emphasizes that the assistant's capabilities on a mobile device have always been and remain strictly limited by the operating system's security mechanisms. Access is possible exclusively to the data and functions for which the user has given explicit permission. According to him, "Alisa" receives no additional capabilities or advantages over foreign models, and its functionality is reduced to basic assistant actions. Moreover, the company representative insists: the gadget owner always has the right to choose an assistant — Russian or foreign — and this fundamental right must remain inviolable.
The essence of the initiative and the risks
According to information circulating in the industry, the draft resolution under discussion could grant "Alisa" the status of a "system assistant" with access to system functions, applications, notifications, and the device's hardware capabilities. It is precisely this wording that has caused a wave of concern. Banking applications store a colossal amount of confidential data: from account balances and CVV codes to passport details. Transactions are conducted through them, and loan applications are submitted. Granting an AI assistant access to this sphere without the strictest restrictions is a colossal risk to citizens' financial security.
It is telling that the draft resolution itself, which appeared on the public discussion portal in early August 2026, was suddenly removed from it. Apparently, the document is currently at the revision stage, which indirectly confirms the complexity and ambiguity of the norms embedded in it.
My view of the situation
Such initiatives always balance on the edge between technological progress and digital security. On the one hand, the integration of AI into everyday scenarios is inevitable and convenient. On the other hand, granting one corporation, even a system-forming one, expanded access to users' financial data without ultra-transparent regulation is a dangerous precedent. The market needs clear rules of the game that exclude any possibility of abuse, not loud statements about security that cannot be verified externally. The question of what exactly will be available to "Alisa" inside banking utilities remains open, and this is perhaps the most alarming signal.