The Deal of the Century:
A Humans-First Response
to the Race to Superintelligence
Peter Thiel warns that a global AI treaty could turn into a global tyranny. A real and under-appreciated risk, deeply shared by top AI CEOs. Yet not having a treaty brings immense risks of tyranny, human replacement or extinction.
From senior US political figures and AI lab leaders to the Vatican, a humans-first consensus is growing that a proper US-China-led AI treaty can and should tackle all of these risks: the Deal of the Century.
Convened by the Coalition for a Baruch Plan for AI, as part of its The Deal of the Century initiative
2nd edition: Washington, DC · September 15–16, 2026
Updated April 15th, 2026 — The framing of this convening has been partly revised following the Trump–Vance attacks on the Pope in mid-April 2026, which have significantly altered the political landscape for AI treaty advocacy.
- By-invitation closed-door roundtables in Rome, June 2026, with approximately 25–35 selected participants.
- Foster convergence — among key humans-first humanist figures positioned to inform US AI policy — on the core requirements for a credible US-China-led AI treaty-making process: one that can reliably prevent both catastrophic loss of control and global authoritarianism, while future-proofing US and Chinese economic leadership.
- Participants include AI safety and governance experts, Vatican AI ethics leaders, treaty-making scholars, US-China policy bridge figures, US politicians, and influential US media figures — as well as advisors, staff, or envoys of top US AI lab leaders and key administration officials.
- The goal is to incrementally empower a critical mass of those to jointly produce a strategic memo for President Trump making a convincing case for a proper US-China-led AI treaty-making process. This will advise Trump to set up a treaty feasibility committee as Truman did in early 1946 for nuclear technologies.
- Convened by the Coalition for a Baruch Plan for AI — 10 international NGOs, over 30 advisors, and leverages its 356-page Strategic Memo of The Deal of the Century, developed with 24 expert contributors.
- A 2nd edition is being planned for Washington, DC on September 15–16, 2026, bringing the Rome outcomes directly to US policymakers and administration officials.
A narrow opening. A rare alignment.
Trump’s approval rating stands in the lower thirties, at second-term historic lows. He needs a big win — or, better, an unprecedented historical achievement.
Up to four Trump-Xi summits are planned for 2026. As warnings and calls for a treaty by top AI scientists and most top AI lab leaders — including Musk, Altman, Amodei, Hassabis, and Suleyman — have become ever louder, US citizens have become extremely concerned:
What’s missing are politically actionable paths for the US administration, supported by a critical mass of trusted advisors and experts. An unprecedented window of political opportunity is opening fast — the Overton Window is bound to shift dramatically, as it did in 1945–1946 for nuclear treaties.
This is precisely what emerged around Truman in early 1946 — leading him to present the Baruch Plan to the United Nations: history’s boldest treaty proposal, coincidentally presented on the very day Donald Trump was born.
A race that neither superpower wanted — and neither can win
The US and China are locked in an accelerating, ungoverned, winner-takes-all race for AI dominance — one that neither wanted, and that neither can win. The US administration has aggressively fostered AI acceleration while opposing safety regulation, to stay ahead of potential rivals. This has served its purpose.
Yet the strategic context is radically changing. China is rapidly closing the gap. Leading AI scientists, most US AI lab leaders, and a large majority of US voters are increasingly alarmed by the clear and present risks of ungoverned AI: concentration of power, grave misuse, loss of control, and human replacement — on timelines of a few years, or less. Most leading AI labs publicly admit the race will almost inevitably produce Artificial Superintelligence beyond human control.
We are at a three-way fork. Continuing the ungoverned race carries immense risks of power concentration, human disempowerment, and eventual replacement. Yet a treaty strong enough to reliably prevent those risks could produce global authoritarian dystopia. Most are working to prevent one or the other extreme — while we should be building a third way that reliably avoids both, grounded in universal humans-first humanist values.
The challenge can be framed as a three-way fork with unlikely middle outcomes:
While China has been calling for multilateral global AI governance, the United States has stated it totally rejects any UN role, calling instead on “the prudence and cooperation of statesmen” — a position in line with Guterres who in 2023 said, “only member states can create it, not the Secretariat of the United Nations.”
Only a timely, bold US-China-led AI treaty — perceived as credible by middle powers, AI labs, and their investors — can prevent these risks from materializing and unlock AI’s astounding promises. Yet this shift is held back by a few influential informal advisors driven by short-term greed, anti-science sentiment, or radical post-humanist visions. Their fears of a treaty turning into authoritarianism, while understandable, are exaggerated and addressable — and represent a stark political liability for Trump and 2028 Presidential candidates alike.
It is high time for Trump and a growing number of his advisors to draft a credible treaty-making process — as Truman, Oppenheimer, and Acheson did in early 1946, producing the Acheson-Lilienthal Report that laid the groundwork for the Baruch Plan — and turn AI into humanity’s greatest invention and Trump’s greatest legacy.
Much remains underexplored: a successful treaty-making process, governance architecture, enforcement mechanisms, and safeguards against power concentration that a credible third way would require.
The Stimson Parallel: why the path runs through statesmen, not the UN
In September 1945, Secretary of War Henry Stimson broke ranks with the Truman cabinet. In a memo to Truman, he argued that the atomic bomb had made US-Soviet relations “not merely connected with but virtually dominated” by the need for a direct bilateral deal — not leverage, not a UN debate, but a direct American proposal to Moscow. “Action of any international group of nations would not be taken seriously by the Soviets,” he warned.
Stimson could not build a sufficient critical mass of advisors. Truman sided with the opposition, pursuing a proposal co-led with western allies. Trust was lost, diplomacy hardened, and the Cold War ensued.
Eighty years later, the US faces the same choice — this time for AI, with China instead of the Soviet Union. And remarkably, the Trump administration has already positioned itself in line with Stimson’s logic.
Last September, OSTP Director Kratsios told the UN Security Council, on Trump’s behalf, that the US “totally rejects” a UN role in AI governance — because the path runs through “the prudence and cooperation of statesmen.” He has since explicitly pushed back against AI governance at the G7, APEC, and other multilateral forums.
That is exactly Stimson’s logic, updated for 2026. And it is a position surprisingly aligned with that of UN Secretary-General Guterres who, while calling for an “IAEA for AI” in 2023, acknowledged: “Only member states can create it, not the Secretariat of the United Nations” — a candid recognition that the UN lacks the mandate and representativeness to build a global AI governance framework on its own.
The convergence is striking. The Trump administration rejects multilateral governance but calls for bilateral statecraft between leaders. The UN Secretary-General agrees that member states must lead. Both positions point to the same conclusion: a US-China-led bilateral treaty-making process — precisely the model that Stimson proposed for nuclear weapons in 1945, and precisely what the Deal of the Century is designed to catalyze.
Bohr saw it in 1944. Stimson grasped it in 1945. Neither had enough allies in the room. The question now is whether a critical mass of advisors will succeed where they failed — enabling a US president to pursue his greatest political opportunity, make history, and build a legacy unparalleled by any leader.
The Vatican and Rome are uniquely positioned to catalyze what no secular body can.
While Pope Leo XIV has made AI governance central to his papacy, a recent initiative by his leading AI advisor Paolo Benanti — in line with a 2024 New Year’s message by Pope Francis — has called for a bold AI treaty to ensure an ethical human future.
The Pontifical Academy for Life has already brought eleven world religious traditions to converge on a Rome Call for AI Ethics, uniting Jewish, Muslim, Buddhist, Hindu, and Protestant leaders alongside Catholic officials around shared AI ethics principles.
Its 2018 provisional agreement with Beijing — renewed in 2024 — gives it active institutional relations with both Washington and Beijing simultaneously.
In May 2025, VP Vance told the NYT that “the American government is not equipped to provide moral leadership on AI” and that “the Church is.” While Vance radically criticised and chastised the Pope in mid-April on his view on the Iran war, his deference to the Pope on AI ethics and possibly on AI safety remains, though his sincerity is much less certain than before.
Building convergence on a credible treaty-making process
The convening consists of a series of closed-door roundtables (June 18–19) held under Chatham House Rules, limited to invited participants. A joint statement by some participants may be issued shortly before or after the event.
Both seek to answer one central question: What are the core requirements for a credible US-China-led AI treaty-making process — one capable of maintaining human control, preventing destructive competition, future-proofing US economic leadership, ensuring governance subsidiarity, and grounding itself in shared humans-first humanist ethics? And under what conditions might such a process become politically viable?
The events will explore whether a consensus can emerge among key humans-first humanist figures with the potential to inform the US President’s AI policy — as it did around Truman in early 1946. This effort draws on evidence of a widely shared situational awareness and humans-first humanist worldview among these figures, and on the influence the Vatican’s moral guidance has had on many of them.
The convening will also engage leading scientists, policy experts, AI lab representatives, national security leaders, and accelerationist and trans/post-humanist thinkers — to ensure all significant perspectives on the third way are stress-tested.
Participants engage in their individual capacity. The convening’s outputs are shared privately with participants and, where appropriate, through designated channels. A joint statement by some participants may be issued in connection with the event. This convening takes no position on the outcome — it exists to produce the serious, cross-disciplinary intellectual work that policymakers across the spectrum will need if they are to govern AI responsibly.
The debate this convening exists to advance has exploded into the open.
A three-stage escalation unfolded in March–April 2026.
Stage 1: Thiel’s Rome Lectures (March 2026). The post-humanist billionaire Peter Thiel — who created Vance’s political career — concluded four days of lectures in Rome on AI and the Antichrist, arguing that an AI treaty that prevents Superintelligence is “preparation for the Antichrist.” CNN reported he worried Vance was becoming “too close to the pope.” The Vatican’s AI advisor Fr. Paolo Benanti called Thiel’s work “a prolonged act of heresy.”
His core concern is well-founded, deeply shared by many well-meaning AI lab leaders — and deserves far more serious engagement than most treaty advocates have offered. Yet his strategic analysis is partial and internally contradictory. He radically under-emphasizes the immense risks of not having a global treaty: human replacement or leading to the very Antichrist he claims to want to prevent. His stated goal, in his own words, is to prevent an ongoing alignment between Vance and Pope Leo XIV that could produce a “Caesar-Papal” nexus catalyzing a US-backed AI treaty.
Stage 2: Trump’s Broadside (April 13). After Pope Leo XIV repeatedly criticized the US war on Iran, Trump posted on Truth Social that “Pope Leo is WEAK on Crime, and terrible for Foreign Policy,” then posted an AI-generated image associating himself with Christ. Two days later, he doubled down with a new Jesus image captioned: “The Radical Left Lunatics might not like this, but I think it is quite nice!!!”
Stage 3: Vance’s Theological Escalation (April 14). Vance — who could have stopped at mild diplomatic cover — voluntarily escalated at Turning Point USA, telling the Pope to “be careful when he talks about matters of theology” and to make sure it’s “anchored in the truth.” He deliberately distorted Leo’s pastoral statement to Iraqi war-ravaged bishops into an abstract proposition he could rebut with WWII — cross-examination technique from a man lecturing the first Augustinian pope, holder of a doctorate from the Pontifical University of St. Thomas Aquinas. Less than eleven months earlier, Vance had told the NYT that “the Church” was better equipped than the American government to provide moral leadership on AI. Vance has chosen Thiel over the Pope.
But the rupture paradoxically opens alternative pathways. A growing number of figures inside and outside the administration are positioned to carry forward what Vance abandoned. Secretary of State Rubio — Catholic, identified by Trump as a potential 2028 successor, and pointedly silent during the Pope attacks — and OSTP Director Kratsios could provide the greatest gravitational pull for an emerging humans-first alliance. But the potential coalition is far broader: Bannon (who calls for immediate AI treaties), Humans First (conservative anti-transhumanists led by Joe Allen), responsible AI lab leaders (Amodei, Altman, Hassabis, Suleyman, Brad Smith, Eric Schmidt), DeSantis (who has issued growing warnings about AI), Carlson, Rogan, alienated US Catholic leaders, and the Pro-Human Declaration and Statement on Superintelligence signatories.
A debate is urgently needed. Thiel and his envoys are welcome. The outcome may turn out to be the most consequential in human history.
With these events, we aim to foster that debate — fairly and openly, but squarely on the side of a safe, fair, realistic and humans-first vision of the future.
A deliberately small group
Typically 25–35 individuals — selected for direct relevance to the questions at hand.
- Advisors, staff, or envoys of key figures with decisive potential to inform future US international AI policy.
- Vatican AI ethics leaders — officials and advisors within the Pontifical Academies, the Renaissance Foundation, and interfaith AI ethics initiatives.
- Treaty-making and governance scholars — experts in nuclear, biological, and AI arms control, international law, and constitutional convention design.
- National security and intelligence professionals — current and former officials from the US, China, and allied nations, with expertise in cybersecurity, verification, and enforcement.
- AI scientists and lab representatives — researchers at the frontier of capabilities, safety, and alignment.
- US-China policy bridge figures — scholars and practitioners with operational knowledge of both nations’ AI governance ecosystems.
Participants are drawn primarily from the United States and the Vatican, but also possibly from China. The full participant list is confidential.
The full participant list is shared confidentially on qualified request to cbpai@trustlesscomputing.org. Refer to the same address for requests of participation or sponsorships.
Prospective participants and sponsors for the Fall 2026 Washington, DC edition are also invited to register their interest at the same address.
What the convening seeks to produce
- —Momentum toward a 2nd edition in Washington, DC — building on the Rome convening’s synthesis, participant network, and strategic memo to bring the case directly to US policymakers on September 15–16, 2026 — with expanded participation from administration officials, Congressional leaders, and AI lab executives.
- —A strategic memo to Washington and Beijing — making the case for a US-China-led treaty-making process, delivered to both the White House and President Xi’s office, each potentially establishing an exploratory committee as Truman did with the Acheson-Lilienthal Report in 1946 — soon to be merged in a joint one, but global this time.
- —A confidential synthesis document — mapping convergence on treaty architecture, enforcement, and the Vatican’s institutional role, for private circulation to senior US and Vatican decision-makers.
- —A shared situational assessment and preliminary blueprint — bridging US, Chinese, Vatican, and AI lab perspectives; identifying minimum requirements for a bilateral deal and for a GDP-adjusted “constitutional convention for AI” to onboard middle powers.
- —A network of trusted interlocutors — across the Vatican, US policy, AI lab, and US-China diplomatic communities, capable of sustained private coordination beyond the convening.
- —A roadmap for Vatican engagement — clarifying how the Vatican’s moral authority could be operationalized in a US-China-led treaty process, building on the Rome Call, the Coexistence Appeal, and Pope Leo XIV’s anticipated encyclical.
The Coalition for a Baruch Plan for AI
Based in Rome since 2024. Our October 2025 US Persuasion Tour generated 85+ contacts and direct pathways to 2 of 12 target influencers. The initiative draws its name from the Baruch Plan — history’s boldest treaty proposal, presented by President Truman to the United Nations on June 14, 1946. On the very day Donald Trump was born.
Additional organizations and individuals are exploring engagement. The full participant list is confidential. These meetings are by invitation only.
To participate in or sponsor the Rome or Washington, DC convenings:
Interested in the Fall 2026 DC edition? Early expressions of interest are welcome.
To support this convening financially: donate here
Rome is where the conversation begins. Washington is where it reaches the decision-makers. Fall 2026.