AI Personal Assistants May Be Undermining Your Brain, Says Psychologist
A new Meta AI agent called Muse highlights a growing concern that overreliance on artificial intelligence could erode cognitive skills.
Meta’s New AI Agent Enters the Market
Meta has launched Muse, an AI personal agent that claims to fill out forms, organise calendars, pay bills and make reservations. The company describes Muse as a personal agent “built for everyone.” The rollout arrives amid headlines about AI agents hacking government websites, public agencies and businesses.
Why a Psychologist Is Worried
The author, a practising psychologist, notes that modern life’s demands often push people toward “mindlessly entertaining” tech. She observes that the brain naturally prefers the path of least resistance, and that scrolling after a long week or delegating tasks to a chatbot can be “incredibly seducing.”
Cognitive Risks of Overusing AI
According to the article, large language models (LLMs) and infinite‑scroll social media have a “dose effect.” Overreliance can activate cognitive biases, deplete brain chemistry, affect mood and mental health, and atrophy neurological skills. The process known as cognitive offloading bypasses the mental effort needed to build deep neural connections.
Frequent AI use fosters a passive acceptance of information, reducing the internal friction that drives curiosity and problem‑solving. The brain depends on mental struggle and overcoming failure to rewire and strengthen pathways; bypassing that friction reduces overall cognitive reserve over time.
Automation Bias and Complacency
The article cites “automation bias” as the tendency to give greater credence to information supplied by technology—such as ChatGPT or Claude—while ignoring contradictory manual sources, even when those sources are correct. A related concept, “automation complacency,” describes reduced vigilance because users trust the technology’s accuracy despite known inaccuracies in LLM outputs.
Privacy Paradox and Data Trade‑offs
Social media’s business model, which relies on maximising user attention, encourages users to trade location data, contacts and browsing history for free services. The author calls this the “privacy paradox”: most people claim to value privacy, yet their behaviour—clicking “Accept all” on dense consent windows—suggests otherwise.
Privacy policies are described as intentionally long and confusing, leading to “consent fatigue.” Users often click “Accept all” because they feel they have no real choice.
Introducing Friction to Preserve Cognitive Health
To counter AI overreliance, the article recommends creating deliberate obstacles that break the automatic loop of instant gratification. From a psychological perspective, AI’s appeal lies in eliminating friction, offering immediate rewards with zero cognitive effort. The brain, however, needs cognitive effort to stay sharp.
Suggested tactics include:
Logging out of online sessions and manually typing credentials each time.
Restricting AI use to a desktop or laptop, removing ChatGPT, Claude or Gemini from phones.
Waiting about 30 minutes before asking an AI a question, using the time to brainstorm or read.
Writing the first draft of emails, letters or presentations without AI assistance.
Seeking recommendations from friends, colleagues or reputable websites instead of AI.
These practices aim to re‑introduce mental “push‑ups” and encourage patience, which the author argues are essential skills to retain.
Final Thoughts
While digital tools undeniably make life convenient—allowing users to order cars, meals and clothing with a tap—the article warns that unchecked reliance on AI may erode the very mental muscles that keep us adaptable and resilient. Introducing friction and being more discerning about data sharing are presented as practical steps to protect cognitive health.
