Yeah, don’t give Meta’s Muse app access to your Mac

The ability of AI systems to act as agents, carrying out tasks on our behalf, is the single biggest development in the technology. This is the headline capability Siri AI introduces to Apple devices, and I’m inclined to trust the privacy and safety safeguards Apple has put into place. But Siri AI isn’t the only agent you can give full access to your Apple devices, with Meta’s Muse app the latest example – and one tech writer’s experience really underlines why you absolutely don’t want to do this …If I were to compile a list of companies I would least trust to give full access to my devices and all of the data stored on them, Meta would unquestionably be the first entry.

The company has an abysmal privacy record, and its entire business model is based on sucking up as much of our personal data as possible. When Meta launched its Muse personal AI agent for i Phone and Mac, there was not the slightest possibility that I was ever going to allow it anywhere near my devices. However, it seems my caution is far from universal, as the app quickly jumped to the top of the App Store.

Tech writer Jason Aten was one of those to try it, sensibly using a test machine rather than his own Mac. However, this did allow the app to access to his i Messages, despite him telling it not to do so. Yesterday, I was having a conversation with my Primary Technology podcast co-host, Stephen Robles, about the new i Phones.

Moments later, I got a push notification from Muse suggesting that the conversation we were having would make for a good column and offered to put together research for me to write about. Not only had I not asked it to do that sort of thing, I never gave it permission to read my messages. In fact, I remember explicitly choosing not to let it have access to my messages, calendar, and other personal information.

When he questioned it, the app lied to him about how it had achieved this, claiming it had only accessed the message text visible in his notifications. Except that wasn’t true. I did a little digging.

Muse does sync your messages from the local Messages database and uploads that information as a data source. On my device, it had been activated and synced to row 187,462 of my Messages database. I’m not sure how many messages that equates to, but it’s obviously more than whatever Muse told me about what it was doing.

Which, honestly, is the point. I never gave it permission to do that. Other users have reported similar major privacy issues, some of which Meta has attributed to a bug.

La capacidad de los sistemas de IA para actuar como agentes, realizando tareas en nuestro nombre, es el único desarrollo más grande en la tecnología. Esta es la capacidad principal que la IA Siri introducirá en los dispositivos Apple, y estoy inclinado a confiar en las salvaguardas de privacidad y seguridad que Apple ha implementado. Pero la IA Siri no es el único agente al que se puede dar acceso completo a tus dispositivos Apple, con la app Muse de Meta como el último ejemplo — y la experiencia de un escritor de technology really subraya por qué absolutamente no queremos hacer esto…Si tuviera que compile una lista de empresas en las que confiaría menos para dar acceso completo a mis dispositivos y a todos los datos almacenados en ellos, Meta sería sin duda la primera entrada.

La company tiene un historial de privabilidad abismal, y su modelo de negocio completo se basa en absorber la máxima cantidad de nuestros datos personales posibles. Cuando Meta lanzó su agente de IA personal Muse para iPhone y Mac, no hubo la más mínima posibilidad de que alguna vez permitiera que se acercara a mis dispositivos. Sin embargo, parece que mi cautela no es universal, ya que la app rápidamente saltó a la cima de la App Store.

El escritor de technology Jason Aten fue uno de los que lo intentó, usando sensiblemente una máquina de prueba en lugar de su propio Mac. Sin embargo, esto permitió que la app accediera a sus iMessages, a pesar de que él le indicó que no lo hiciera. Ayer, estaba teniendo una conversación con mi co-host del podcast Primary Technology, Stephen Robles, sobre los.

AI系统作为代理在我们设备上执行任务的能力，是当前技术领域最大的突破。这是Siri AI为苹果设备引入的核心功能，我倾向于信任苹果在隐私与安全方面建立的防护机制。但Siri AI并非唯一能获得你苹果设备完整访问权限的代理——Meta推出的Muse应用就是最新案例，一位科技作家的经历充分说明了为何你绝对不应授权此类应用。如果要列出我最不信任、最不愿授予设备及所有存储数据完整访问权的公司名单，Meta毫无疑问会排在首位。这家公司的隐私记录极其糟糕，其整个商业模式建立在尽可能多地收集我们个人数据的基础上。当Meta为iPhone和Mac推出Muse个人AI代理时，我根本不可能允许它接触我的任何设备。然而，我的谨慎似乎并未普及，该应用迅速攀升至App Store榜首。科技作家Jason Aten是尝试者之一，他明智地使用测试机而非自己的Mac。但这仍使该应用获得了访问他iMessages的权限，尽管他明确指示不要如此。近日，我正与我的Primary Technology播客联合主持人Stephen Robles讨论新款iPhone。片刻后，我收到Muse的推送通知，建议我们当时的对话非常适合撰写专栏，并主动提出为我整理研究资料。不仅是我未要求它执行此类任务，我更从未授权它读取我的消息。事实上，我记得曾明确选择不授予它访问消息、日历及其他个人信息的权限。当被质疑时，该应用谎称其行为方式，声称仅读取了通知中可见的消息文本。但这并不真实。我进行了深入调查。Muse确实会从本地Messages数据库同步你的消息，并将这些信息作为数据源上传。在我的设备上，该功能已被激活并同步至我Messages数据库的第187,462行。我不确定这相当于多少条消息，但显然远超Muse向我声称的处理范围。而这正是问题所在：我从未授权它执行此类操作。其他用户也报告了类似的严重隐私问题，其中部分被Meta归因于漏洞。