Artificial Intelligence Policy Tracker
Search
Menu

AI incident ·

Inappropriate Gmail Smart Reply Suggestions

22 news reports Snapshot 7 Sep 2026
On this page 9 sections

In brief

An AI system built and deployed by Google allegedly harmed Gmail Users.

Risk domain
Human-Computer Interaction Overreliance and unsafe use
Occurred
Coverage
22 reportsNov 2015 - Oct 2018

What happened

Google's Gmail Smart Reply tool was over-recommending the response "I love you" in situations where it was deemed innappropriate.

Laws on record for this risk area

Policy angle: Classified under Human-Computer Interaction (Overreliance and unsafe use) in the MIT AI Risk Repository taxonomy; 5 instruments on record cover this risk area.

Matched from the record's risk category and country to the instruments in effect recorded here: a law listed covers the area, which is not a finding that it applies to this incident. A reviewer can correct the match in the repository (data/external/incident_overrides.yaml).

News reports (22)

Titles link to the original publisher; report text is not reproduced here.

  1. Computer, respond to this email: Introducing Smart Reply in Inbox by Gmail
    blog.google · Bálint Miklós
  2. New Gmail feature scans your emails and suggests replies
    money.cnn.com · David Goldman
  3. Soon, Gmail's AI Could Reply to Your Email for You
    wired.com · Cade Metz
  4. Google Inbox Suggests Replies for Incoming Email
    techlicious.com · Elmer Montejo
  5. Google's Smart Reply Lets a Computer Answer Your Emails for You
    popularmechanics.com · Eric Limer
  6. Google Inbox's new Smart Replies uses machine learning to speed up email responses
    greenbot.com · Blair Hanley Frank
  7. Google's new email feature will write replies for you
    mashable.com · Karissa Bell
  8. Google’s New Autoreply Sounds Great!!!!
    newyorker.com · Nicola Twilley, Isaac Chotiner, Aimee Lucido
  9. Google's Email Replying AI Keeps Saying It Loves You
    escapistmagazine.com · Fanghawk
  10. Google's Smart Reply: Convenience Over Connection
    thetechnoskeptic.com · Samantha Maxwell
  11. Ranking Gmail’s AI Fueled Smart Replies
    nymag.com · Christopher Bonanos
  12. Gmail's Smart Reply Blurs the Line Between People and Brands
    dailydot.com · Brenden Gallagher
  13. Is This Article Worth Reading? Gmail’s Suggested Reply: ‘Haha, Thanks!’
    wsj.com · David Marcelis, Douglas Macmillan
  14. Gmail users will be able to opt out of Smart Replies on desktop
    engadget.com · Mallory Locklear
  15. Smart Reply Not Smart Enough? Desktop Gmail Users Can Soon Opt Out
    digitaltrends.com · Chuong Nguyen
  16. Gmail Smart Reply goofed w/ 'Sent from my iPhone'
    9to5google.com · Stephen Hall
  17. You can turn off Gmail's creepy Smart Replies on desktop too - just not yet
    trustedreviews.com · Chris Smith
  18. Neira Jones on Twitter
    twitter.com · Neira Jones
  19. Google tweaked its Smart Reply feature because it would repeatedly suggest users send 'I love you'
    dailymail.co.uk · Annie Palmer For Dailymail.Com
  20. Google changed Gmail's Smart Reply after AI kept suggesting I love you
    businessinsider.com · Kif Leswing
  21. Gmail smart compose: Will it change how we write emails?
    qz.com · Corinne Purtill
  22. Google ends support for Reply, which added smart replies to third-party apps
    theverge.com · Natt Garun

Who was involved

Alleged deployer
Google
Alleged developer
Google
Alleged harmed party
Gmail Users

Classification (MIT AI Risk Repository taxonomy)

Causal entity
AI
Intent
Unintentional
Timing
Pre-deployment
Harm level
none
Sectors
information and communication
Countries
—

Risk entries describing this failure mode

Entries from the MIT AI Risk Repository coded to subdomain 5.1.

  • Overreliance

    "Over-reliance - Unfettered and/or obsessive belief in the accuracy or other quality of a technology system, resulting in addiction, anxiety, introversion, sentience, complacency, lack of critical thinking and other actu...

    A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms (Abercrombie2024)

  • Addiction

    "Addiction - Emotional or material dependence on technology or a technology system."

    A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms (Abercrombie2024)

  • Alienation/isolation

    "Alienation/isolation - An individual’s or group’s feeling of lack of connection with those around as a result of technology use or misuse."

    A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms (Abercrombie2024)

  • Economic instability

    "Economic instability - Uncontrolled fluctuations impacting the financial system, or parts thereof, due to the use or misuse of a technology system, or set of systems."

    A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms (Abercrombie2024)

  • Alteration of social relationships may induce psychological distress

    The Rise of Artificial Intelligence - Future Outlooks and Emerging Risks (Allianz2018)

  • Social manipulation in elderly- and child-care

    " the use of advanced AI for elderly- and child-care are subject to risk of psychological manipulation and misjudgment "

    The Rise of Artificial Intelligence - Future Outlooks and Emerging Risks (Allianz2018)

  • Healthcare

    "the use of advanced AI for elderly- and child-care are subject to risk of psychological manipulation and misjudgment (see page 17). In addition, concerns about patients’ privacy when AI uses medical records to research...

    The Rise of Artificial Intelligence - Future Outlooks and Emerging Risks (Allianz2018)

  • Inconsistent Performance across and within Domains

    "Estimating true capabilities of an LLM is a difficult task (c.f. Section 3.3), especially for naive users unfamiliar with the brittle nature of machine learning technologies. Exaggeration of model capabilities by the de...

    Foundational Challenges in Assuring Alignment and Safety of Large Language Models (Anwar2024)

Incidents in the same risk subdomain

All incidents in this subdomain

Other incidents involving Google

Source record: incident #17 on the AI Incident Database · all 22 reports