AAAI-27: Dates, Location, Deadlines & Registration Guide

TL;DR: AAAI-27 runs February 16-23, 2027 at the Palais des Congrès de Montréal in Montréal, Canada. The 41st AAAI Conference on Artificial Intelligence is a peer-reviewed research event, not a vendor floor, with a 17.6% acceptance rate at AAAI-26 on 23,680 submissions (checked 2026-08-24, csconfstats.xoveexu.com). Registration pricing and keynote speakers are still unpublished. The full paper deadline is July 28, 2026.

I track AI events because what ships in production tools six months from now usually got a paper accepted at a conference like this first. AAAI-27 is not that conference in the sense most guides frame it. There is no exhibitor hall to walk, no vendor lounge, no keynote sizzle reel. A paper’s acceptance letter matters more than any invited speaker AAAI eventually names, and forums quoting an AAAI-27 acceptance rate right now are guessing at a number AAAI has not published yet.

Why AAAI-27 is a submission decision, not a travel decision

AAAI-27 is the 41st edition of a genuinely selective conference. AAAI-26 in Singapore drew 23,680 Main Track submissions and accepted 4,167, a 17.6% acceptance rate (checked 2026-08-24, csconfstats.xoveexu.com/conferences/aaai/2026/). That is down from 23.4% the year before as submission volume nearly doubled. The Main Track sits alongside EAAI-27 (education), IAAI-27 (applied), a Doctoral Consortium, New Faculty Highlights, a Senior Member track, and a Special Track on AI for Social Impact.

The point of that number is not the number itself. It is what the November 30, 2026 decision does to a researcher’s calendar. Paper clears, you attend. Paper does not clear, you do not. That is the actual decision most attendees are making. Not “is this conference worth a trip.”

When AAAI-27 runs and what the venue actually asks of you

AAAI-27 runs February 16-23, 2027 at the Palais des Congrès de Montréal in Montréal, Canada (checked 2026-08-24, aaai.org/conference/aaai/aaai-27/). The eight-day window covers workshops, tutorials, and bridge programs early, then the main technical program with EAAI-27 and IAAI-27 through the core days. Montréal is a rational venue: direct international routes through Montréal-Trudeau, enough hotel inventory for a conference regularly clearing 20,000-plus submissions, and a local research community (Mila, McGill, Université de Montréal) with genuine skin in the field.

Practical note. Mid-to-late February in Montréal is real winter. Budget for it, both in clothing and in flight-delay buffer.

AAAI-27 at a glance:

Field	Value
Event	AAAI-27 (41st AAAI Conference on AI)
Dates	February 16-23, 2027
Venue	Palais des Congrès de Montréal, Canada
Format	Primarily in-person; virtual TBA
Full paper deadline	July 28, 2026 (UTC-12)
Final decisions	November 30, 2026
Registration	Not yet published for AAAI-27
Official site	aaai.org/conference/aaai/aaai-27/

The submission timeline you plan around

AAAI-27’s review cycle runs on a publicly posted schedule, all deadlines UTC-12 (verified against aaai.org/conference/aaai/aaai-27/submission-instructions/, July 2026). Author registration for the Main Technical Track opened June 17, 2026. Paper submission opened June 30, 2026. Abstracts are due July 21, 2026, full papers July 28, 2026, supplementary materials and code by July 31, 2026. Phase 1 rejections land September 24, 2026. Author feedback runs October 19-25, 2026. Final decisions arrive November 30, 2026. Camera-ready is due December 14, 2026.

Two format traps get papers desk-rejected before a reviewer opens them. The first is the style file. Papers must use the AAAI two-column style from the official AAAI-27 Author Kit, which contains aaai27.sty, aaai27.bst, and the anonymous submission template for double-blind review. Grabbing an older aaai2027.sty from a GitHub mirror is the most common way to get rejected on format. Format for US Letter, embed Type 1 or TrueType fonts only, and do not alter the margins or two-column layout aaai27.sty defines.

The second trap is reciprocal reviewing. AAAI-27 expects all authors of a submission to join the reviewer pool unless extenuating circumstances prevent it. Ignoring the nomination step can put your own submission at risk.

The AI-assisted peer review pilot AAAI is running again

AAAI is continuing the AI-Assisted Peer Review Pilot into AAAI-27 after running it in AAAI-26. Per the official Main Technical Track call for papers, Phase 1 review pairs two standard human reviews with one AI-generated, non-decisional review. The AI review carries no rating or recommendation. It only becomes visible to the assigned human reviewers, area chairs, and Program Committee members after the human reviewers have submitted their own reviews, during the paper discussion phase, not before. AAAI is explicit that no paper content is used for LLM training, and humans alone make every accept or reject decision.

That is a meaningfully more cautious position than “let AI review your paper.” It treats the AI output as a second opinion inside the discussion phase, not a gatekeeper. Worth knowing before you submit, since your paper’s AI review will exist even if it never influences the outcome.

What AAAI-27 registration will probably cost

AAAI has not published AAAI-27 pricing. Any article printing exact 2027 figures now is guessing. The honest reference point is AAAI-26’s Singapore pricing (per aaai.org/conference/aaai/aaai-26/registration/, checked 2026-08-24).

Registration type	Current member (early)	New/renewing (early)
Regular, in-person	$750	$870
Student, in-person	$350	$390
Virtual	$80	n/a
Student virtual	$40	n/a

Two things worth knowing before AAAI-27 pricing lands. Student pricing sat around half the regular member rate, so if you are a graduate student and your advisor is not funding travel, budget accordingly. Virtual attendance at $80 is roughly 10% of the in-person price, so if the blocker is travel budget rather than research quality, virtual is a real option, assuming AAAI offers it again. Whatever AAAI-27 ultimately charges, expect the same tiered shape: cheaper before the early-bird cutoff, cheaper for students, cheaper for virtual.

Who this conference serves and who it does not

AAAI-27 is built for people doing original AI research. The core audience is PhD students and academic researchers submitting papers, faculty and postdocs tracking research trends, and a smaller but growing group of applied-AI research leads at labs like Google DeepMind, Meta FAIR, and Anthropic who attend as an early signal for which techniques are about to become production-relevant.

You get real value from AAAI-27 if you fit one of these profiles.

• Presenting a paper, poster, or Doctoral Consortium submission
• Faculty or postdoc shaping the next research direction for your lab
• Industry research lead scouting for techniques 12 to 18 months ahead of the tools market
• Already embedded in the review process as a reviewer or area chair

If you are not submitting a paper and not doing hands-on AI research, this is the wrong conference. There is no vendor floor, no product demo track, no beginner “learn about AI for buyers” programming. Wandering an academic poster session without a research background feels like reading a paper in a language you only half speak. If you buy and evaluate AI tools rather than research them, the AI tools I’ve tested is the direct signal you want.

My buy-or-skip framing for AAAI-27. Submit first if you are a researcher, then decide travel after the November 30, 2026 decision. Attend virtually if the format is offered and travel is the blocker. Skip entirely if your interest in AI is on the buying side. Watch the official AAAI-27 page directly for pricing and keynote confirmations rather than any third-party list, since speaker rumors on academic Twitter run months ahead of AAAI’s actual announcements. For a broader research-plus-industry audience alternative, ICML 2026 and NeurIPS 2026 sit closer to the research and industry boundary than AAAI does.