SI vs AI: What the US Rename Means for Hospitality

By Ankur Handoo, Co-founder, HuemanAI · Last updated 1 October 2026 · 9-minute read

Key Takeaways

AI is the field. Superintelligence (SI, also written ASI) is a hypothetical level within it. SI would outperform humans across virtually every cognitive domain.
SI has not been demonstrated. The US government's new "SI" label, set by an executive order on 29 September 2026, covers the same technologies the law already calls AI. It relabels existing systems; it does not create a new tier.
Hospitality already runs on narrow and generative AI. One 2026 survey found 98% of hoteliers had used AI in the past six months; another found only 32% of owners have it embedded across most operations.
The data points to integration, trust and skills, not raw intelligence, as the bottleneck. Build on what's proven and keep people where guests value them.

On 29 September 2026, the US government directed federal agencies to use "Super Intelligence" (SI) instead of "AI" in official documents. That creates a naming clash for anyone comparing SI vs AI, because SI already has a specific meaning in AI research. This guide explains both meanings of SI, the current evidence on AI in hospitality, and what superintelligence could and couldn't mean for hotels and restaurants. Every number has a date and a source, and where something is a forecast rather than a fact, we say so.

What is artificial intelligence (AI)?

Artificial intelligence is the umbrella term for software that performs tasks normally needing human intelligence: understanding language, spotting patterns, forecasting demand, making recommendations. Everything deployed in hospitality today, from a revenue-management tool to a voice agent taking reservation calls, belongs here. You'll meet three flavours: predictive AI (forecasting and pricing), generative AI (writing, summarising and conversation) and agentic AI (planning and taking actions across tools, such as checking availability and updating a booking).

How capable is it? The 2026 International AI Safety Report, guided by over 100 experts from more than 30 countries, found that leading systems reached gold-medal level on International Mathematical Olympiad problems, yet performance stays uneven: they still fail at some simple tasks and are less reliable across many-step projects. For now, the report concludes, agents complement rather than replace humans.

What is superintelligence (SI)?

Philosopher Nick Bostrom's widely used definition, from his 2014 book Superintelligence, is any intellect that greatly exceeds human cognitive performance in virtually all domains of interest (as cited here). The key phrase is virtually all: a system that surpasses humans in one domain but can't be applied outside it is domain-specific, not generally superintelligent. The usual ladder has three rungs: narrow AI (defined tasks), artificial general intelligence or AGI (matching humans across tasks) and artificial superintelligence, written ASI or SI (exceeding them). IBM calls ASI hypothetical and notes that AGI would be a big step towards it.

Does it exist? No. Fortune reported in November 2025 that no such systems exist and that scientists debate whether today's methods can reach it, though Meta and others have made it a stated goal. The live argument has moved one rung down. In early September 2026, after OpenAI launched GPT-6 Astra, its president Greg Brockman said it's reasonable to feel we're in the AGI era and Nvidia's Jensen Huang said AGI had arrived, but the ARC Prize Foundation, which tested the model, said it wasn't claiming AGI (Information Age). Treat "AGI is here" as contested. For superintelligence, we found no credible demonstration.

How soon? Forecasts differ widely. A large 2023 survey of 2,778 AI researchers gave a 50% aggregate chance that machines outperform humans at every task by 2047, and a 50% chance that all occupations are fully automatable by 2116. Those are opinions, not measurements.

Did the US government rename AI "Super Intelligence"?

Yes, for its own documents. On 29 September 2026, President Trump signed an executive order titled "Inaugurating the Era of Super Intelligence", directing executive-branch agencies, to the maximum extent permitted by law, to use "Super Intelligence" and "SI" instead of "Artificial Intelligence" and "AI" in official correspondence, websites, reports and other non-statutory documents. He previewed the change at the UN General Assembly on 22 September, arguing that "artificial" makes the technology sound unreal (The Hill, Forbes India). The order followed a White House meeting where tech executives signed a voluntary accord (NewsNation).

What it doesn't do. The White House argues that increasingly capable systems make the new name more appropriate, yet the order defines SI as the technologies and systems already covered by the legal definition of AI in 15 U.S.C. §9401(3). It doesn't replace that statutory definition (IAPP), doesn't retroactively alter existing regulations, contracts or grants (The Tribune), and binds executive agencies, not Congress, the courts, the states or companies (HNGN). A proposed federal definition of SI is due within 60 days, around 28 November 2026.

SI now has two meanings:

Dimension	Research meaning	US federal usage
Definition	An intellect that greatly exceeds human cognitive performance in virtually all domains	The same technologies the law already calls AI
Status	Hypothetical	In use for executive-branch non-statutory documents since 29 Sep 2026
Describes	A capability level	A label

Experts have flagged the clash. Marketing experts told Newsweek that confusion with the term's established technical meaning is a major reason it's unlikely to be adopted soon. For hospitality, nothing changes in what hotel or restaurant technology is or does.

SI vs AI: the key differences

Superintelligence isn't the opposite of AI. It's the proposed top rung of it. Here's how the three levels compare.

Capability Level	Narrow AI	AGI	Superintelligence (SI/ASI)
What it does	Excels at defined tasks	Would match humans across most tasks	Would exceed humans in virtually all domains
Status, October 2026	Deployed widely	Claimed by some, disputed by experts	Hypothetical
In hospitality	Voice agents, forecasting, scheduling	No agreed evidence yet	None. Any claim would be speculation

How AI is helping hospitality today

Adoption is broad, depth is not

Mews, a hotel PMS vendor, surveyed 500+ properties between December 2025 and March 2026: 98% of hoteliers had used AI in the previous six months, with AI involved in 11 of 19 common hotel tasks. Wyndham's survey of 325 owners and developers in the US, Canada and the Caribbean, run by Wakefield Research, found 98% had begun incorporating AI, yet only 32% said it was embedded across most operations. In restaurants, the National Restaurant Association's 2026 report found 26% of operators using AI-related tools, with marketing the top use, and only 6% using AI to take customer orders.

The figures differ because having used AI is not the same as having it embedded, so always check the definition. BCG's global study puts hospitality below the cross-industry average, though with only 12 hospitality companies sampled, treat that as directional.

Discovery is moving to AI, booking trust lags

BCG and NYU's Tisch Center found 37% of travellers already use large language models embedded in travel sites to plan and book trips. Hilton (AI Planner, beta, March 2026), Marriott (Ask Bonvoy, June) and IHG (beta, July) have all launched conversational search, and Hilton reported slightly higher conversion among AI Planner users (Skift).

Yet Expedia's YouGov survey of 5,700+ adults in the US, UK and India (March 2026) found only 8% of US and UK travellers relied on chatbots such as ChatGPT or Gemini to plan a trip, and nearly 70% preferred booking with a trusted travel brand. Expedia commissioned that research, and the two studies measure different behaviours. Together they suggest AI is entering discovery faster than it is earning booking trust.

Voice and chat agents

On its Q1 2026 earnings call, Wyndham said more than 1,100 hotels were live on its Wyndham Connect+ platform, each effectively with its own AI voice agent answering calls and chats. Management reported nearly 300 basis points of incremental direct contribution and 25% faster call handling. Those are company-reported figures, not independent audits.

Operations

BCG and NYU cite early deployments with 20% faster room cleaning and preparation from AI-synchronised housekeeping schedules, and roughly 50% less food waste within eight months from AI kitchen analytics.

Where humans still win

In the Mews survey, 59% of hoteliers said front-desk welcome and check-in should stay human-led, most strongly among properties already using AI extensively.

What is the scope of superintelligence in hospitality?

We found no credible evidence that a superintelligent system exists, so no hotel can have deployed one and any claimed result would be unverifiable. The evidence supports a three-tier view:

Tier	What it is	Hospitality evidence
Proven	Narrow and generative AI	Measured gains in call handling, housekeeping and food waste, mostly company- or survey-reported
Emerging	Agentic AI and AI-led discovery	Conversational search at Hilton, Marriott and IHG; IHG is also in a Google agentic-booking pilot. Gartner predicts over 40% of agentic AI projects will be cancelled by end-2027 over cost, unclear value or weak risk controls
Hypothetical	Superintelligence	None. By definition it would beat humans in virtually all domains, but that follows from the definition, not from any hospitality result

Whatever comes next, the data shows where hotels are stuck today. BCG reports nearly half of hoteliers struggle to access critical information across patchwork systems, and only 2.9% of full-time travel and tourism employees have AI skills, versus 21% in tech and media. Mews found 41% of hoteliers have no formal AI policy, and properties with one report 92% strong trust in their AI tools versus 49% without. Read together, these figures suggest the limit on hospitality AI today is integration, governance and skills rather than raw intelligence.

What hospitality leaders should do now

Write an AI policy covering data use, guest disclosure and escalation to humans.
Fix data and integrations first. Connect PMS integration, reservations and CRM before adding autonomous agents.
Start with bounded, measurable use cases, such as call handling and reservations, and set KPIs before launch. Ask vendors for evidence: Gartner warns many rebrand chatbots as agentic AI.
Keep people in high-value moments such as arrival and complaint handling.
Disclose AI. Since 2 August 2026, EU AI Act Article 50 has required AI systems that interact directly with people to tell them so, unless that's obvious. If your AI talks to guests in the EU, check your obligations.
Make your content machine-readable. BCG and NYU list high-trust, machine-readable content as the first priority for staying visible in AI recommendations.
Watch the terminology. If a vendor calls its product "SI", ask whether it means the research definition or the new US label.

Where HuemanAI fits. We build AI agents for hotels and restaurants, including an AI receptionist and voice agents, an AI concierge and restaurant table management, designed to connect with PMS and reservation platforms such as Oracle Opera, Mews, Guesty and Eveve.

The bottom line

AI is what hotels and restaurants are using today. Superintelligence, in the research sense, is a hypothetical level nobody has demonstrated, whatever label a government document uses. The evidence to date suggests the practical work matters most: connected data, clear policies, measurable use cases, and people kept in the moments guests value. If you want AI agents that connect to your PMS and handle guest calls and bookings, book a demo with HuemanAI.

Bridge the Gap Between Proven AI and Real Revenue

Deploy autonomous AI voice agents and guest communication layers that integrate seamlessly into your PMS—capturing every direct booking 24/7 without operational overhead.

Book a Demo

FAQ

What is the main difference between superintelligence and AI?

AI is the broad field of machines performing tasks that normally need human intelligence. Superintelligence (SI) is a hypothetical level of AI that would exceed human cognitive performance in virtually all domains. Today's deployed systems are useful but uneven.

Did the US government rename AI to superintelligence?

Yes, for its own documents. A 29 September 2026 executive order directs executive-branch agencies to use "Super Intelligence" and "SI" instead of "AI" in non-statutory documents, and defines them as the same technologies the law already calls AI. In AI research, SI still means a hypothetical level that exceeds humans in virtually all domains.

Does superintelligence exist yet?

No. Companies have stated it as a goal, and as of October 2026 even AGI, a lower bar, is disputed. The US "SI" rename covers the same systems as before, so it doesn't change that answer.

Is ChatGPT superintelligence?

No. The International AI Safety Report 2026 describes leading systems as strong on some very hard problems yet failing at some simple ones, so they don't outperform humans across virtually all domains.

How is AI used in hotels today?

Common uses include voice and chat agents for guest enquiries, conversational search on brand sites, housekeeping scheduling, revenue management and kitchen waste tracking. Mews found AI involved in 11 of 19 common hotel tasks.

Will AI replace hospitality staff?

The current evidence points to augmentation: 59% of surveyed hoteliers want check-in to stay human-led, and the International AI Safety Report finds agents complement rather than replace humans for now. Wyndham's management says its voice agents take labour out of front offices (a company claim).

Sources and further reading

Definitions and the SI debate

The US "Super Intelligence" executive order

Hospitality evidence

Governance

Superintelligence (SI) vs AI: The Difference, the US Rename and What It Means for Hospitality