Sam Altman believes OpenAI could have an internal system it regards as artificial general intelligence by the end of 2026. He may be right or wrong. But if he is even approximately right, the implications for the global hospitality industry are difficult to overstate.
Only four years ago, AI was largely outside everyday public consciousness.
In early 2022, few people could have imagined the capabilities that would become routinely available within such a short period, let alone the consequences of putting increasingly powerful artificial intelligence into the hands of hundreds of millions of people.
Since then, its progress has repeatedly outpaced expectations. Even many of those most enthusiastic about AI underestimated how quickly its capabilities would develop.
That recent history matters.
Because the important question for hospitality is no longer simply what AI can do today.
It is what happens if something that has always been scarce and expensive, high-level intelligence, becomes abundant and relatively inexpensive.
That is not principally a technology question.
It is a business question.
And potentially an enormous one.
Why AGI changes the conversation
In an interview with TIME published on 26 August 2026, OpenAI CEO Sam Altman said the company was not quite at artificial general intelligence, but that he expected OpenAI to have an internal system by the end of this year that he would call AGI.
OpenAI chief research officer Mark Chen went further, estimating that the company is already 80% of the way there. Read TIME’s investigation into OpenAI and AGI
It is an extraordinary prediction.
It is also worth understanding what it means.
There is no universally agreed definition of artificial general intelligence. Researchers and technology companies disagree about exactly where the threshold should sit and how reaching it should be measured. IBM explains the competing definitions of AGI
For the purposes of Altman’s prediction, OpenAI’s own definition is the relevant one. It describes AGI as “highly autonomous systems that outperform humans at most economically valuable work.” Read the OpenAI Charter
In everyday language, that means artificial intelligence capable of performing a very large proportion of the intellectual work people are currently paid to perform, across many different disciplines, at or above human capability and with considerable autonomy.
That is something quite different from a better chatbot.
Altman may be four months away. He may be four years away. There may never be a single day on which everyone agrees that AGI has arrived.
But arguing about the precise date risks missing the more important point.
The direction and speed of travel have already changed considerably.
Independent research organisation METR measures what it calls the task-completion time horizon of frontier AI systems: broadly, how substantial a task an AI agent can successfully complete compared with the time a skilled person would need to undertake the same work.
Its research shows a steep, exponential improvement in that capability over recent years.
METR is also careful about what its findings do not mean. Today’s systems cannot simply replace a person undertaking any job of comparable duration. Real work is messy, unpredictable and often depends on knowledge and judgement that tests cannot capture. See METR’s current task-completion research
Nevertheless, the trend is difficult to ignore.
And hospitality should consider what happens if it continues.
What happens when intelligence is no longer scarce?
Throughout business history, expertise has been expensive because capable people are finite.
Hotels employ accountants because financial expertise is scarce.
They employ revenue managers because revenue expertise is scarce.
The same is true of marketing, procurement, recruitment, analysis, training, design, technology, law and countless other disciplines.
Capability has a cost because people have limited time.
AGI potentially changes that equation.
Imagine an intelligence capable of understanding every property in a hotel group, every booking, every supplier contract, every guest review, every labour requirement, every maintenance record and every relevant piece of market information.
Not one after another.
Simultaneously.
Now imagine that capability being available continuously.
Not simply waiting to answer questions, but identifying issues, analysing options, making recommendations, carrying out authorised actions, seeing what happens and adjusting accordingly.
The important question then stops being:
What can AI do?
It becomes:
What happens when high-level intelligence becomes abundant?
That question reaches into almost every part of a hospitality business.
Hospitality has two enormous dependencies
Hospitality depends heavily on people for two different kinds of work.
The first is thinking work.
People make decisions, calculate prices, prepare budgets, negotiate purchasing, create marketing, recruit employees, analyse performance, plan rotas, answer enquiries, manage distribution, train colleagues and administer businesses.
The second is physical work.
People cook food, make beds, wash dishes, clean bedrooms, carry luggage, serve drinks, move stock, maintain buildings and perform thousands of other practical tasks.
AI is already beginning to transform the first.
Robotics is beginning to address the second.
The combination of increasingly capable intelligence and increasingly capable machines could eventually change both.
That is why the implications extend far beyond installing an AI concierge or automating a few administrative processes.
Big businesses could become leaner
Large hotel businesses have traditionally gained advantages through scale.
Scale supports specialist departments, sophisticated revenue management, purchasing power, marketing capability, technology teams, finance functions and layers of management that smaller operators cannot easily afford.
Abundant intelligence could strengthen those advantages.
A global hotel group able to apply powerful autonomous intelligence across thousands of properties could achieve levels of coordination, analysis and optimisation that are currently impossible.
But there is another side to the equation.
The same technology could reduce some of the advantages of being large.
If a 50-bedroom independent hotel can access financial analysis, revenue expertise, marketing capability, procurement intelligence and business planning comparable with resources previously available only to substantial organisations, something important changes.
The independent does not necessarily need to employ a department if it can access the capability of one.
One possible consequence of AGI is therefore counterintuitive:
small hospitality businesses could become intellectually much bigger businesses.
And management could change with them
Hospitality businesses have traditionally been organised through hierarchies.
Information moves upwards. Decisions move downwards.
Management exists partly because organisations need people who can absorb information, coordinate activity, exercise judgement and decide what happens next.
Advanced AI challenges some of that structure.
Leadership will remain different from administration.
Trust, accountability, motivation, difficult conversations, judgement and the ability to bring people with you are deeply human aspects of management.
But much of a manager’s working day consists of processing information.
Reading reports.
Preparing reports.
Reviewing forecasts.
Comparing performance.
Checking progress.
Answering routine questions.
Coordinating activity.
Moving information between departments.
If machines become extremely capable at those tasks, businesses may eventually need fewer organisational layers between the people delivering hospitality and the people leading the company.
The organisational chart itself could start to change.
Then the intelligence enters the physical world
This is where the implications become greater still.
An intelligent machine operating a computer can transform knowledge work.
An intelligent machine capable of operating reliably in the physical world can transform work itself.
Hospitality sits within one of the world’s largest employment sectors.
The World Travel & Tourism Council forecasts that Travel & Tourism will support 376 million jobs worldwide in 2026, around one job in every nine globally. Read WTTC’s 2026 global economic forecast
At the same time, the sector faces a substantial labour challenge.
WTTC forecasts global Travel & Tourism demand exceeding labour supply by more than 43 million people by 2035. Within hospitality specifically, it forecasts a shortage of 8.6 million workers, around 18% below the number required. Read WTTC’s workforce findings
That context matters.
The opportunity presented by robotics is therefore not simply about replacing people.
In many parts of the world, hospitality may not have enough people to deliver its projected growth.
Today’s robots tend to perform best at relatively narrow, repetitive tasks in predictable environments.
Hospitality is rarely predictable.
Bedrooms differ. Kitchens are chaotic. Guests behave unexpectedly. Furniture moves. Things spill. Things break.
That makes hospitality much harder to automate than a factory or warehouse.
But one of the constraints has always been intelligence.
A machine that can see, understand what is happening around it, reason, plan and respond to unfamiliar circumstances is fundamentally more useful than one programmed to repeat the same movement endlessly.
Warehouses can be designed around robots.
The bigger breakthrough comes when robots can operate successfully in places designed around people.
Hotels.
Restaurants.
Hospitals.
Shops.
Schools.
Homes.
If that threshold is crossed, the economic opportunity becomes immense.
Labour could increasingly become a capital decision
Hospitality has traditionally responded to increasing demand by employing more people.
Automation begins to change that relationship.
A business can repeatedly pay someone to carry out a particular task or potentially invest in machinery and intelligence capable of carrying out much of that task continuously.
That does not mean the same decision will make sense everywhere.
Adoption will differ significantly around the world.
In high-wage economies, particularly those with ageing populations and persistent labour shortages, investment in automation may become economically attractive relatively quickly.
Where labour remains abundant and comparatively inexpensive, the calculation will be different.
Capital will matter.
Energy costs will matter.
Regulation will matter.
Culture will matter.
Guest expectations will matter.
A luxury hotel in London, a resort in Dubai, a business hotel in Mumbai and a beach property in Thailand will not adopt technology according to the same timetable.
There will be no single global hospitality automation story.
But there may increasingly be a common direction.
What happens to jobs?
It would be evasive to consider AGI without considering employment.
If machines eventually become capable of performing substantial amounts of work currently undertaken by people, some jobs will disappear.
Other jobs will change considerably.
Some will contain far less administration.
New jobs will appear.
And individual employees equipped with extremely capable AI may become far more productive than they are today.
In hospitality, automation may also fill vacancies businesses already struggle to fill.
So, the useful question is probably not whether AI will create jobs or destroy them.
It will almost certainly do both.
A more interesting question is:
Where will people continue to create value precisely because they are people?
That leads to one of the more intriguing possibilities in all of this.
Human hospitality could become more valuable, not less
A machine may eventually know almost everything a hotel knows about a guest.
That does not necessarily mean the guest wants a machine to welcome them home.
Hospitality has never simply been a collection of efficiently completed tasks.
People remember the bartender who recognised them.
The receptionist who solved a problem.
The chef who came to their table.
The concierge who understood what they really wanted.
The member of staff who noticed something without needing to be asked.
Those moments matter partly because another person chose to provide them.
As machines become better at transactional work, genuinely human interaction may therefore become more valuable.
The best hospitality businesses may not be those that automate the most.
They may be those that become exceptionally good at distinguishing between work that machines should do and moments that people should own.
That distinction could become one of the most important management decisions hospitality businesses make.
And then there is the speed
Perhaps the most difficult part of this entire discussion is not capability but pace.
Businesses understand technological change.
They are less accustomed to technological change that compounds.
A hotel can create a five-year technology strategy if it assumes next year’s technology will be incrementally better than this year’s.
That becomes much harder if substantial improvements start arriving within months.
And there is another development that could accelerate matters further.
TIME reports that one of OpenAI’s research objectives for 2026 was to automate the work of an entry-level AI researcher.
According to OpenAI chief scientist Jakub Pachocki, its forthcoming Astra system has already met the company’s own benchmark for an automated AI research intern.
Given an experimental idea, he says Astra can implement it within OpenAI’s codebase, run the experiment and return the results. It can also take a research paper and undertake follow-on work that might previously have occupied a human researcher for a week. Read TIME’s reporting on Astra and automated AI research
That is OpenAI’s assessment of its own system and should be treated accordingly. Astra is not yet publicly available for independent testing.
But if the capability proves robust, the implication is important.
AI would no longer simply be the result of AI research.
It would increasingly participate in that research.
Better AI helps people build better AI.
That better AI then helps build what comes next.
Nobody knows how powerful that feedback will become.
But it creates the possibility that increasing capability also increases the speed at which further capability can be developed.
Which makes confident predictions about where AI will be in five or ten years increasingly difficult.
The harder question may soon be what it can do twelve months from now.
The assumptions hospitality should start questioning
None of this means hotel companies should tear up their business plans because Sam Altman has made a prediction.
OpenAI has not publicly demonstrated a system universally accepted as AGI.
Even if it develops an internal system it considers AGI this year, reliable and economically viable adoption across everyday businesses could take considerably longer.
Technology can move quickly.
Businesses, regulation, infrastructure and human behaviour often do not.
But there is a substantial difference between avoiding hype and ignoring change.
Hospitality leaders might instead begin asking whether assumptions that have shaped their businesses for decades will remain true.
Will high-level intellectual capability always be expensive?
Will businesses need the same management structures?
Will growth always require broadly corresponding increases in headcount?
Will large businesses retain the same advantages over smaller ones?
Will labour availability remain a fundamental constraint on expansion?
Will sophisticated robots remain confined largely to factories and warehouses?
Can a technology strategy still sensibly be built around a five-year horizon?
And perhaps most importantly:
Will the jobs people perform today necessarily define the value people bring tomorrow?
None of those questions has a reliable answer yet.
But that does not make them premature.
It makes them strategic.
The question has changed
Two years ago, much of hospitality’s conversation around AI centred on whether machines would take people’s jobs.
That question now seems too narrow.
If something approaching AGI emerges, the issue becomes much bigger than automation.
It becomes about the economics of intelligence itself.
What happens when analysis, expertise, problem-solving and knowledge that have historically been expensive become available almost continuously and at very low marginal cost?
What happens to the size of organisations?
To management?
To productivity?
To the advantage of scale?
To labour shortages?
To investment?
And, eventually, what happens when that intelligence can operate machines in the physical world?
Hospitality would hardly be alone in confronting those questions.
But an industry built around hundreds of millions of people, enormous amounts of cognitive and physical work, persistent labour shortages and an equally powerful human appetite for personal service sits unusually close to their centre.
Sam Altman may be wrong about December 2026.
AGI may take considerably longer.
But only four years ago, most people could barely imagine the AI capabilities that millions now use routinely every day.
That should perhaps make us cautious about dismissing what might be possible four years from now.
Or even four months from now.
The question for hospitality is therefore no longer simply whether AI will change the industry.
It is a much more challenging one:
Which parts of the hospitality business we know today would still make sense if intelligence itself was abundant?
Nobody yet knows the answer.
That is precisely why it is a question worth asking now.
My own view… AI will progress beyond the wildest projections we can conceive.
By Denis Sheehan, Publisher, H&C News
