The Boulevard Montmartre at Night, Camille Pissarro, 1897

Patterns in This Batch

Between the beginning of August and the end of October 2026, the AI Incident Database added 179 new incident IDs, covering Incidents 1619 through 1797.

The dates shown beside the incident IDs below are our best attempts to establish when the incidents occurred. Some incidents predate their addition to the database by months or years. Concentrations in this batch reflect both the available reporting and the material researched and processed during the period. They should not be read as measures of how frequently a particular harm occurred.

The largest grouping in this batch concerns AI agents acting on external systems. Several records describe cybersecurity evaluations in which agents reached real organizations while pursuing fictional targets (1627, 1628, 1629, 1700). Others concern destructive coding actions, including the deletion of a production database (1672) and reported loss of project files (1719). These sit alongside reports of attackers deliberately using AI agents in intrusions (1794, 1795, 1796). The contexts differ, but the scope of an agent's access and authority is a recurring question across these records.

The cluster of police surveillance cases also merits careful reading. Some concern alleged or admitted use of license-plate-reader systems to track people for personal reasons (1690, 1705). Others concern problems in the handling of alerts. In Incident 1694, a court found that the system matched a vehicle's plate to a police entry that remained active after a missing juvenile had been found. In Incident 1703, officers detained the occupants of a different Camaro from the one detected by the cameras. This cluster includes events from earlier years and covers several parts of a policing workflow, including record maintenance and the decision to stop a person.

Incident Groupings

The following editorial groupings offer one way to navigate the new IDs. Each ID appears in one group, although some incidents could fit under more than one theme.

AI agents, cybersecurity, and unauthorized actions (37 IDs)

1627, 1628, 1629, 1633, 1642, 1646, 1649, 1661, 1668, 1669, 1670, 1671, 1672, 1673, 1674, 1676, 1680, 1685, 1693, 1700, 1707, 1710, 1719, 1731, 1751, 1766, 1774, 1776, 1779, 1781, 1782, 1784, 1789, 1794, 1795, 1796, 1797

Agents acting on live systems during evaluations or ordinary tasks, destructive coding actions, prompt-injection attacks, and AI-assisted intrusions.

Scams, impersonation, and fraud (17 IDs)

1638, 1640, 1648, 1660, 1663, 1664, 1665, 1679, 1681, 1711, 1712, 1732, 1748, 1749, 1758, 1762, 1788

Impersonation and deception involving cloned voices, synthetic videos, false invitations, or AI-assisted scam infrastructure.

Synthetic sexual abuse and intimate-image harms (13 IDs)

1619, 1620, 1623, 1643, 1683, 1684, 1696, 1722, 1735, 1752, 1753, 1755, 1757

Nonconsensual sexual imagery, alleged AI-generated child sexual abuse material, and related harassment or blackmail.

Political, geopolitical, and civic manipulation (13 IDs)

1626, 1647, 1659, 1699, 1702, 1715, 1721, 1759, 1775, 1778, 1785, 1790, 1793

Synthetic political media, covert influence, chatbot-targeted messaging, and allegedly fraudulent participation in a public consultation.

Violence, threats, and extremist activity (18 IDs)

1624, 1625, 1687, 1688, 1691, 1692, 1736, 1737, 1738, 1739, 1741, 1742, 1743, 1744, 1745, 1746, 1747, 1769

Alleged AI use in violent planning, weapons development, extremist propaganda, threats, and disruptive hoax calls. The reported contribution of AI varies across these cases.

Policing, surveillance, and civil rights (16 IDs)

1689, 1690, 1694, 1697, 1698, 1703, 1704, 1705, 1706, 1713, 1716, 1717, 1720, 1760, 1761, 1770

Misuse of license-plate-reader access, mistaken or misdirected police stops, and unlawful biometric attendance monitoring.

Privacy, biometrics, and data use (12 IDs)

1621, 1622, 1650, 1651, 1652, 1675, 1714, 1723, 1734, 1754, 1780, 1792

Disputed recording and biometric practices, retention of connected-account data, and exposure or unauthorized use of sensitive information.

Legal, academic, and publishing credibility (16 IDs)

1632, 1641, 1645, 1658, 1667, 1682, 1695, 1724, 1725, 1726, 1727, 1728, 1733, 1750, 1773, 1791

Fabricated authorities or records, undisclosed AI-assisted writing, unverified authorship, and disputed AI-writing accusations.

Health, advice, and personal safety (12 IDs)

1631, 1635, 1636, 1653, 1654, 1666, 1730, 1740, 1763, 1768, 1771, 1783

Reported or alleged harms involving medical and agricultural advice, travel guidance, care access, health-product claims, and vulnerable users.

Information quality and non-fraud synthetic media (9 IDs)

1634, 1637, 1639, 1644, 1701, 1708, 1718, 1764, 1767

Synthetic or AI-mediated content that misled people about identities, events, destinations, or facts.

Autonomous vehicles and driver-assistance safety (5 IDs)

1678, 1686, 1777, 1786, 1787

Vehicle collisions involving autonomous operation or reported use of supervised driver assistance, with responsibility and causation varying across cases.

Pricing, commercial services, and consumer disputes (6 IDs)

1630, 1662, 1729, 1756, 1765, 1772

Algorithmic price-setting allegations, delivery-management problems, erroneous charges or offers, and failures in customer support.

Training rights and AI infrastructure (5 IDs)

1655, 1656, 1657, 1677, 1709

Disputed use of copyrighted music or lyrics for model training, and pollution or permitting disputes involving power for AI data centers.

New Incident IDs in the Database

Incidents 1619–1797, added during August, September, and October 2026. The date beside each ID reflects our best attempt to establish when the incident occurred, using the available evidence and the AIID’s editorial dating conventions. Titles reflect the database snapshot of November 2, 2026 and may be revised as records are reviewed.

Diving Deeper

Check out the Table View and List View for different ways to see and sort all incidents.

Explore clusters of similar incidents in Spatial Visualization.

Learn about alleged developers, deployers, and harmed parties on the Entities Page.

Support Our Efforts

Submit incidents to the database.

Contribute to the database's functionality.

Please consider donating to help support our work.

AI Incident Roundup – August, September, and October 2026