The AIS ETA field has no year. The US Coast Guard’s specification page lists it as 20 bits: month, day, hour, minute, in UTC, with a default for “not available” (NAVCEN, Class A reports). That one field says a lot about AIS vessel tracking. The dot on your map and the arrival date next to it travel in different messages, and nothing in the message itself checks the second one.
Two broadcasts with different owners
A Class A transponder sends a position report every 2 to 10 seconds while underway and every 3 minutes at anchor. Voyage data (destination, ETA, draught) goes out in a separate message every 6 minutes (NAVCEN AIS FAQ). The position comes from the ship’s electronic position-fixing system, and the NAVCEN Class A page says it is understood that the overwhelming majority of those sensors are GPS. The position report even carries a flag for exceptions: a time-stamp value of 62 means the fix is dead reckoning, an estimate, and 61 means manual input.
The voyage message has no such sensor. Rynda’s AIS reference says destination is free text (“RTM”, “ROTTERDAM”, “FOR ORDERS”), draught is often the last voyage’s figure, and the ETA is often stale because the crew types it in. It puts AIS navigational status in the same category, a field the officer of the watch sets by hand, and says an anchored ship showing “under way” is common. A tracker that prints “arriving Thursday” straight from that ETA field is quoting whatever was last entered.
So AIS ship tracking accuracy splits in two. Where the ship was a few seconds ago is a measurement. When it will be alongside a berth is, per Rynda, whatever the crew last typed in. NAVCEN’s FAQ tells operators to check the accuracy of their AIS data before each voyage and says US-waters rules require accurate input and upkeep of the data parameters, so the duty exists and rests on the crew, and the broadcast itself does not show whether it was done.
A silent dot is usually a receiver problem
Terrestrial receivers hear a ship out to roughly 20 to 40 nautical miles ship to ship, more to a high shore antenna, and beyond that satellites pick it up on a pass, minutes to hours later (Rynda). A page from Sanctioned Vessels adds that commercial platforms stitch the feeds together and interpolate, so a tidy line across an ocean is partly drawn. Rynda’s rule of thumb is that a two-hour gap mid-Atlantic is routine and a two-hour gap off a busy coast is not.
A single gap does not say which case it is. Judge it by duration, by location and by a second source, and read every position together with its age.
AIS spoofing: what it changes on the map and what it leaves alone
NAVCEN describes AIS as an open, unencrypted radio system that can technically be spoofed, and its advice is “trust, but verify.” GPS spoofing visible in AIS data has been measured at scale. A 2026 preprint on arXiv ran motion analysis over AIS data from more than 367,000 vessels between late November 2024 and early February 2025 and found 31 persistent anomaly hotspots, at least 22 with strong evidence of GPS spoofing. The authors found the same pattern in the Red Sea near Sudan five months before the MSC Antonia grounded there, and say their measurements are a lower bound on spoofing activity.
Jamming is a different failure: it blocks or degrades the signal, so a receiver is left with a missing or unreliable position, where spoofing hands it a false one. Al-Monitor, quoting AFP, reported in March 2026 that since the US-Israeli strikes began, jamming had left about 1,000 ships in the Gulf and the Gulf of Oman unable to determine their location, “momentarily or continuously.” A Kpler analyst put that at about half the vessels in the area. Todd Humphreys of the University of Texas told AFP that many ships only listen to the original civilian GPS signal, and AFP concludes that most cannot rely on Galileo or BeiDou when GPS is jammed.
What a bad fix does not touch is the cargo record. Panvaya, a tracking vendor, points out that container milestones such as gate moves, loading confirmations and discharge scans come from terminals and carriers and are generally separate from the ship’s live position. A container loaded on a named ship at a recorded port stays loaded on it while the dot jumps. Panvaya’s advice for a claimed diversion is to ask the carrier for its latest operational message. A screenshot of the same AIS feed proves nothing new. That is a vendor’s view. Panvaya also notes that milestones can arrive late, be corrected or be mapped differently by different carriers.
Which date to hang a truck on
For a shipper the rule is simple. Dray bookings, warehouse slots and customer promises should hang on events a terminal or carrier recorded: discharge, gate-out availability, a change in the carrier’s published vessel schedule. In AIS vessel tracking the typed ETA is useful as a hint. Speed over ground in the position report is a better hint than the typed arrival date, though it comes from the same GPS receiver, so in an area with reported interference it needs a second look too.
Navo’s container tracking homepage says it has recorded 110,000+ AIS positions on 2,700+ vessels (it dates the vessel count to its production database, July 2026) and says its container events are cross-checked across two to three sources with 0 to 5 minutes of freshness. The cross-check is the figure worth asking any tracker for, ours included: which sources agreed before this status went on the screen? A vessel position that nothing else corroborates is one input to the decision. Carrier milestones have their own problems, since Panvaya notes they can be late or corrected, which is another reason to want two sources to agree.