Anxiety wants the first answer. Give it a second one.

When a decision makes you anxious, the quickest relief is to grab the nearest answer, often someone else’s. Letting the doubt come back once or twice usually gets you somewhere better. Then stop: past a couple of rounds, more doubt turns into going in circles.
The rush to settle
Anxiety is uncomfortable, and an answer, any answer, ends it. In negotiation experiments, anxious people made weaker first offers, answered faster and walked away sooner (Brooks & Schweitzer, 2011, one study). People who feel unsure also lean harder on advice (Bonaccio & Dalal, 2006, a review).
That is what makes a ready-made answer so tempting when you are worried, including one from an AI. People tend to weigh advice from an algorithm more heavily than the same advice from a person (Logg, Minson & Moore, 2019, one lab).
Write your own first
No study has tested this directly for personal decisions, but the nearby evidence points one way. Making your own call before seeing an AI’s suggestion reduced over-reliance on it (Buçinca, Malaya & Gajos, 2021; Fogliato et al., 2022, single studies). Seeing examples early narrows what people go on to think of (Jansson & Smith, 1991, repeated).
So write your own options and your own lean before you ask anyone, or anything, for theirs. That is why Sleep On It asks for your options first, and offers suggestions only when you ask.
Let the doubt return, twice
Deliberately considering the opposite, or any alternative, reduces the pull of your first view (Lord, Lepper & Preston, 1984; Hirt & Markman, 1995, repeated). But dose matters. In one study, asking people for a long list of counterarguments made their bias worse, because finding so many felt hard (Sanna, Schwarz & Stocker, 2002).
So ask for one: the best case against your lean. Write it down. Let the doubt come back a second time if it wants to. Then leave it.
Then stop
Past a few rounds, doubt stops adding information and starts eroding your willingness to commit to anything (Ward et al., 2003). That is the reason to set a time. The time keeps the question open past the anxious first answer, and it is also where the doubting ends.
Two outside views that help
Ask what you would tell a friend in the same spot. People reason more wisely about other people’s problems than their own (Grossmann & Kross, 2014). And once you have written your own lean, ask one person who has actually done the thing. In one study, a single report from someone who had lived it predicted people’s enjoyment better than their own imagining did (Gilbert et al., 2009).
More questions
Should I ask an AI what to do?
Write your own options and lean first, then ask. People tend to over-weight algorithmic advice, and deciding first reduces that.
Sources
- Brooks AW, Schweitzer ME (2011). Can Nervous Nelly negotiate? OBHDP 115(1).
- Bonaccio S, Dalal RS (2006). Advice taking and decision-making. OBHDP 101(2).
- Logg JM, Minson JA, Moore DA (2019). Algorithm appreciation. OBHDP 151.
- Buçinca Z, Malaya MB, Gajos KZ (2021). To trust or to think. Proceedings of the ACM on HCI 5.
- Fogliato R et al. (2022). Who goes first? FAccT.
- Jansson DG, Smith SM (1991). Design fixation. Design Studies 12(1).
- Lord CG, Lepper MR, Preston E (1984). Considering the opposite. Journal of Personality and Social Psychology 47(6).
- Hirt ER, Markman KD (1995). Multiple explanation. Journal of Personality and Social Psychology 69(6).
- Sanna LJ, Schwarz N, Stocker SL (2002). When debiasing backfires. JEP: Learning, Memory, and Cognition 28(3).
- Ward A et al. (2003). Can’t quite commit. Personality and Social Psychology Bulletin 29(1).
- Grossmann I, Kross E (2014). Exploring Solomon’s paradox. Psychological Science 25(8).
- Gilbert DT et al. (2009). The surprising power of neighborly advice. Science 323.