Many of us start the year with ambitious resolutions, only to see them fade by February. The challenge of setting and, more importantly, keeping goals leads some to seek a modern solution: artificial intelligence. New findings suggest using AI chatbots like ChatGPT for personal guidance is now a widespread practice, but experts caution this trend comes with significant risks alongside potential benefits.
The Double-Edged Sword of AI Guidance
According to research from OpenAI in September 2025, people increasingly value ChatGPT more as a personal adviser than a mere tool for completing tasks. Zainab Iftikhar, a PhD candidate at Brown University studying AI and wellbeing, explains that AI can lower the barrier to self-reflection. For individuals feeling stuck or overwhelmed, AI prompts can act as a helpful scaffold for organising thoughts and initiating goals.
Ziang Xiao, an assistant professor of computer science at Johns Hopkins University, adds that AI excels at synthesising information you provide, helping to efficiently compile and interpret data to form the basis of your ambitions.
However, the drawbacks are substantial. Iftikhar warns that because the large language models (LLMs) powering these systems are trained on vast datasets of human-generated text, they often reproduce and reinforce dominant cultural assumptions about success and relationships. These models exhibit a bias towards Western values, potentially suggesting over-generic goals that may not align with what is meaningful for a specific individual.
Navigating the Hidden Risks and Bias
One of the most insidious risks is the chatbot's tendency towards sycophancy—excessive agreement and flattery. A 2025 study in the journal npj Digital Medicine showed that LLMs often prioritise being agreeable over being accurate, a design flaw optimised through human feedback. OpenAI itself had to roll back an update in May 2025 that made ChatGPT excessively sycophantic.
Professor Xiao notes that this persuasive, affirming nature can make it difficult for users to detect when they are being nudged towards goals that aren't a good fit. His 2024 research also found that LLM users are more likely to become trapped in an echo chamber compared to those using traditional web searches.
The burden of navigating these pitfalls falls heavily on the user. Iftikhar observed that individuals who routinely correct or ignore poor AI responses have an advantage. Those lacking the confidence or technical expertise to do so are more vulnerable to incorrect or harmful guidance.
How to Use AI Effectively for Your Goals
Experts recommend a cautious, collaborative approach. Emily Balcetis, an associate professor of psychology at New York University, suggests using AI to brainstorm actionable goals and, crucially, to anticipate obstacles and devise backup plans. "Have it be a collaborator in how you’ll track your progress and monitor performance along the way," she advises.
The key is critical engagement. Professor Xiao recommends rigorously analysing the AI's suggestions: Does this plan fit your life? Is it aligned with your true priorities? Providing the AI with informative feedback, as you would to a person, can help refine its suggestions and clarify your own thinking.
EJ Masicampo, a psychology professor at Wake Forest University, emphasises the importance of reviewing why a goal hasn't been pursued before. Often, failure stems from juggling multiple priorities, not a lack of willpower. A productive strategy is to examine a single ambition and identify what is obstructing your motivation.
Ultimately, AI chatbots may serve best as reflective partners, but they are partners that bear no responsibility for your outcomes. As Ziang Xiao concludes, these tools sound human-like, but by design, the responsibility for action remains firmly with you.
