IJCAI 2026: Dates, Speakers, Program & Registration Guide

IJCAI-ECAI 2026 ran August 15-21, 2026 in Bremen, Germany, jointly with the European Conference on AI. 990 accepted papers landed across the program (713 Main Track, 277 special tracks), split between University of Bremen for workshops August 15-17 and Congress Centrum Bremen for the main conference August 18-21 (checked 2026-08-24, 2026.ijcai.org). Award talks came from Nicholas R. Jennings, David C. Parkes, and Jiajun Wu. Toby Walsh delivered the free public AI Lecture.

I track AI events the same way I track AI tool deals, by checking primary sources before writing anything. For IJCAI-ECAI 2026 that meant cross-referencing the official conference site against Bremen’s press materials and each award recipient’s own institution, not aggregators quoting aggregators. One site I hit while researching this had workshops at a venue that did not match anything on the official schedule. That is how a PhD student books the wrong flight.

IJCAI-ECAI is not a trade show pretending to be a conference. It is a peer-reviewed academic conference where getting a paper accepted matters more than any badge scan. The two-venue split, the award-talk structure, the visa rules, and the deliberately-opaque registration all reflect that.

Bremen’s two-venue split was the sharpest planning choice

IJCAI-ECAI 2026 broke the week deliberately. August 15-17 ran as workshops and tutorials on the University of Bremen campus. August 18-21 ran as the main conference at the Congress Centrum Bremen. That ordering matters: workshops came first, so attendees could go deep on a subfield or a tool before the main track’s dense parallel-session schedule started.

If your grant only covered part of the week, the workshop window was a legitimate standalone trip. Subfield-specific workshops ran self-contained, and a paper still in progress could pull direct feedback without you sitting through four days of parallel oral sessions. A pair of colleagues splitting the week (one for workshops, one for main-track presentation) covered both ends of the conference without doubling the travel budget.

Bremen’s own logistics helped. The city has its own airport with connections through major European hubs and sits on Germany’s ICE rail network, so travelers routing through Frankfurt or Amsterdam did not need a complicated last leg. The Maritim Hotel Bremen connects directly to the CCB main venue, and several other hotels sit within walking distance of Bremen Central Station.

I judge academic AI conferences on whose talks got named

IJCAI-ECAI 2026 did not run a commercial-style keynote stage. Its major talks tied to named research awards, a meaningfully different (and more verifiable) format than an unnamed “industry keynote.”

Nicholas R. Jennings, Vice-Chancellor of Loughborough University, received the 2026 Award for Research Excellence for his contributions to multi-agent systems and agentic AI. His talk, “The Past, Present and Future of Agentic AI,” placed the current agentic AI wave inside a much longer research history.

David C. Parkes, Dean and George F. Colony Professor of Computer Science at Harvard University, received the 2026 John McCarthy Award, confirmed via Harvard’s own announcement, for foundational work at the interface of AI and economics. His talk, “From Intelligent Agents to Agent Societies,” extended his mechanism-design work into how agent societies should be structured.

Jiajun Wu, Assistant Professor of Computer Science and Psychology at Stanford University, received the Computers and Thought Award for contributions across computer vision, generative AI, and robotics.

Separately, Professor Toby Walsh delivered the IJCAI-ECAI 2026 AI Lecture, a free public talk on the opportunities, limits, and societal implications of AI, opened by the University of Bremen’s president. It was aimed at the general public, not the research audience. Admission was free, though registration was required.

You attended IJCAI-ECAI 2026 for a specific paper track or a specific award recipient’s line of work. Not for a lineup built to look impressive on a brochure.

990 papers, and every one got both an oral talk and a poster slot

The main track packed 713 papers. Special tracks packed 277 more across AI and Health, AI and Robotics, AI for Social Good, AI4Tech, and Human-Centred AI. Add Demonstrations, a Journal Track, Sister Conferences Best Papers, and an Early Career Spotlight, and the total came to 990 accepted papers (checked 2026-08-24, 2026.ijcai.org).

Every accepted paper got both an oral talk and a poster session. Posters are where most of the hallway-conversation value happens at a conference this size, so the format rewarded people who worked the poster hall rather than only camping in one oral track.

The special-track mix read the field’s institutional attention accurately. AI and Health explicitly pushed beyond straightforward applications. AI and Robotics sat at the intersection of generative AI and structured reasoning. Human-Centred AI covered social, cultural, and ethico-legal dimensions alongside the technical work. A Survey Track collected synthesis papers rather than only novel results, useful for a state-of-the-field primer on a subarea without needing to read forty narrow results.

I hit the pricing wall at the same place every reader will

Registration ran a three-tier schedule: early (May 15 to June 15), late (June 16 to July 15), and onsite (July 16 through August 21, closing noon CEST). Base delegate and student registration fee amounts never appeared on the indexed pages. Those numbers only showed up inside the official registration system.

What was public up front (checked 2026-08-24, 2026.ijcai.org/registration):

Item	Confirmed
Member discount	$50 USD off any main conference registration
3-year membership	$100 USD regular / $50 USD student
Banquet add-on (August 19)	$150 USD
Opening reception add-on (August 17)	$120 USD
Student reception add-on (August 20)	$90 USD
Payment method	Credit card only

Full delegate registration included the opening reception and banquet tickets and free tutorials. Student registration included the opening and student receptions and free tutorials. If the member versus non-member gap was close to $100, buying a 3-year IJCAI or EurAI membership at registration paid for itself in the same transaction, especially for anyone attending again.

A separate fee is worth flagging so you do not confuse it with attendee registration: IJCAI’s Primary Paper Initiative charged authors $100 USD per non-primary paper submission, funding reviewer compensation. That is an author-side submission fee, not something attendees paid at the door.

There was no StackVerdict discount code for IJCAI-ECAI 2026 and I am not going to invent one.

The no-virtual rule caught more people out than the visa rules did

IJCAI-ECAI 2026 stated plainly that if none of a paper’s authors could attend in person to present, the paper got removed from the conference proceedings. There was no hybrid, no livestream, no proxy read-in. If you had an accepted paper, someone from your team had to physically be in Bremen.

That rule caught more people out than the Schengen visa timelines did. EU nationals and citizens of Australia, Canada, Iceland, Israel, Liechtenstein, New Zealand, Norway, South Korea, Switzerland, and the USA did not need a visa for stays under three months. Everyone else needed a Schengen visa with:

• A completed application form, valid passport, and biometric photo
• Health insurance valid in the Schengen area plus travel insurance of at least €30,000
• Proof of financial means and intent to return (an employment contract or enrollment certificate)
• Outbound and return flight reservations and proof of accommodation
• An invitation letter, downloadable from inside the registration system after completing the visa form under Options

Schengen processing times ran into weeks. Anyone hitting the visa question in July was already cutting it close by the August start. There was also an Industry Day on Wednesday, August 19 (a US$150 add-on ticket), and student authors who could not cover the cost applied to the conference Grants program or the Volunteers’ program.

I keep IJCAI-ECAI on the shortlist, and here is why

For active AI researchers, PhD students with an accepted paper, and applied ML engineers who want direct exposure to peer-reviewed work ahead of the trend cycle, IJCAI-ECAI 2026 was one of the strongest general AI research conferences on the calendar. The scale (990 accepted papers), the joint IJCAI-ECAI research base, and award recipients from Loughborough, Harvard, and Stanford are not small claims.

The honest caveats are about fit. There was no product floor, no vendor keynote, no “state of AI in business” briefing. Enterprise buyers were better served by a commercial AI summit. And the base registration price only appeared inside the process, so budgeting six months out was harder than with a published price list.

The framing that has held up for me: go if you have a paper accepted, a workshop presentation, or a specific research question the program directly addresses. Send a colleague instead if you are evaluating this purely as a networking or business-development event, because that is not what this conference optimizes for. The best AI tools I have tested for research and technical workflows often trace back to IJCAI or NeurIPS papers from 12 to 24 months earlier, so watching the accepted-paper list stays worthwhile long after the badges come off.

IJCAI 2027 has not published dates or a host city yet. The registration and program pages for 2026 stay live on 2026.ijcai.org for anyone who wants to work back through the papers or check the visa page.