AGI/Singularity: 9,800 Predictions Analyzed (original) (raw)

Artificial general intelligence (AGI) is when an AI system matches human cognitive abilities across all tasks. We analyzed 9,800 AI researchers‘, leading entrepreneurs‘, and community predictions about the AGI timeline:

Will AGI/singularity happen? AGI is inevitable according to most AI experts.

When will we reach AGI? Between late 2020s and early 2030s. AGI timeline shortened after ChatGPT launch.

Artificial General Intelligence timeline

Loading Chart

The timeline above outlines the anticipated year of the singularity, based on insights gathered from 10 surveys with over 6,000 participants, responses from 18 AI researchers, and community insights from the prediction markets Manifold, Kalshi, and Metaculus.

As you can see above, survey respondents increasingly expect the singularity to occur earlier than previously anticipated. Learn the methods we used to create this graph.

AGI predictions from prediction markets

For prediction markets & community insight predictions, we used:

Over 3,700 predictions from Manifold, Kalshi, and Metaculus, which are online prediction markets where participants trade on the likelihood and timing of future events for profit or reputation. We also included the average prediction date obtained from these prediction markets.

Other key questions on AGI

What is our current status on AGI?

Although narrow AI surpasses humans in specific tasks, a generally intelligent machine doesn’t exist. Some researchers state that large language models demonstrate emerging generalist capabilities.1 According to our AGI benchmark, machines are far from generating economic value autonomously.

How can we reach AGI?

Either by putting more compute and data behind current architectures like transformers, or by inventing new approaches. There is no scientific consensus yet on the method for achieving AGI or for validating it.

Below you can see a summary of predictions that make up this timeline, or see the complete table.

Results of major surveys of AI researchers

We examined the results of 10 surveys involving more than 6,000 AI researchers and experts, in which they estimated when AGI/singularity might occur.

While predictions vary, most surveys indicate a 50% probability of achieving AGI between 2040 and 2061, with some estimating that superintelligence could follow within a few decades.

Expert Survey on Progress in AI

In October 2023, AI Impacts surveyed 2,778 AI researchers on when AGI might be achieved. This survey included nearly identical questions to the 2022 survey. Based on the results of the Expert Survey on Progress in AI, the high-level machine intelligence is estimated to occur by 2040.2

Expert Survey on Progress in AI

The survey was conducted with 738 experts who published at the 2021 NIPS and ICML conferences. Based on the results of the Expert Survey on Progress in AI, the experts estimate that there’s a 50% chance that high-level machine intelligence will occur by 2059.3

Experts also predicted that hardware cost, algorithmic progress, and work on training sets would be the biggest factors in AI progress.

Forecasting AI Progress Survey

Baobao Zhang surveyed 296 AI experts in 2019, asking them to predict when machines would surpass the median human worker in performing over 90% of economically relevant tasks. According to the results of the Forecasting AI Progress Survey, half of the respondents estimated this would happen before 2060.4

AI Experts’ Survey on AGI Timing

The predictions from the AI Experts’ Survey on AGI Timing in 20195 are:

Survey on AI’s Potential Impact on Labor Displacement

Ross Gruetzemacher surveyed 165 AI experts in 2018 to assess the potential impact of AI on labor displacement. The experts were asked to estimate when AI systems would be capable of performing 99% of tasks for which humans are currently paid, at a level equal to or exceeding that of an average human.

Based on the results of the AI’s Potential Impact on Labor Displacement survey, half of the respondents predicted this milestone would be reached before 2068, while 75% anticipated it would occur within the next 100 years.6

AI experts in the NIPS and ICML conferences survey

In May 2017, 352 AI experts who published at the 2015 NIPS and ICML conferences were surveyed.7

Based on survey results from the NIPS and ICML conferences, experts estimate a 50% chance that AGI will occur by 2060. That said, there’s a significant difference of opinion based on geography:

Some significant job functions expected to be automated by 2030 include call center reps, truck driving, and retail sales.

Future Progress in Artificial Intelligence survey

Vincent C. Muller, the president of the European Association for Cognitive Systems, and Nick Bostrom, from the University of Oxford, who has published over 200 articles on superintelligence and artificial general intelligence (AGI), conducted the Future Progress in Artificial Intelligence survey in 2012 and 2013. 550 participants answered the question: When is AGI likely to happen?8

According to the results of the Future Progress in Artificial Intelligence survey:

Survey with AI experts participating in the AGI-09 conference

Based on the survey results from 21 AI experts participating in the AGI-09 conference in 2009, AGI will occur around 2050, and plausibly sooner.9 You can see below their estimates regarding specific AI achievements: passing the Turing test, passing third grade, accomplishing Nobel-worthy scientific breakthroughs, and achieving superhuman intelligence.

Figure 1: Results from the survey distributed to attendees of the Artificial General Intelligence 2009 (AGI-09) conference.

We also evaluated Samotsvety Forecasting and Metaculus Community predictions on AGI, and prediction market results from Manifold, Kalshi, and Polymarket:

Samotsvety Forecasting

Samotsvety Forecasting is a team of forecasters that makes probabilistic predictions about real-world events, especially in geopolitics, technology, and global risks, using structured reasoning and quantitative methods. They demonstrate a strong competitive track record on major forecasting platforms and tournaments (e.g., INFER/CSET-Foretell), where their accuracy is measured using formal scoring metrics such as the Brier score.10

In January 2026, the team updated their predictions on AGI with 8 forecasters.11 Here are the aggregated results:

In an earlier forecast from 2022, the team estimated a 32% chance of AGI within 20 years (by ~2042) and 73% by 2100, both lower than their current projections.12

Manifold Market

As of April 2026, over 1,100 Manifold market contributors predicted the year when an AI will first pass a “high-quality, adversarial Turing test” as 2033.13

Kalshi Prediction Market

As of April 2026, contributors to the Kalshi prediction market state that there is a 55% chance that OpenAI will achieve AGI by 2030.14

Polymarket

Results from Polymarket predictions in April 2026 indicated that there is a 14% probability that OpenAI achieves AGI by 2027.15

As of April 2026:

In 2022, 81 participants answered the question “When will top forecasters expect the first Artificial General Intelligence to be developed and demonstrated?” and their prediction was 2035.19

Insights from AI entrepreneurs & individual researchers

AI entrepreneurs are also making estimates on when we will reach singularity, and they are more optimistic than researchers. This is expected as they benefit from increased interest in AI.

Their opinions differ on its speed and development pathway. Amodei from Anthropic expects AGI to arrive in the near term due to rapid self-reinforcing progress, while Hassabis from DeepMind sees it as plausible but remains cautious, citing unresolved challenges in scientific creativity and autonomous self-improvement.

Here are the predictions of 15 of the most prominent AI entrepreneurs and researchers:

Other comments and developments about AGI

AAAI Presidential Panel on the Future of AI Research

475 respondents, mainly from academia (67%) and North America (53%), were asked about progress in AI. Though the AAAI 2025 Presidential Panel on the Future of AI Research survey didn’t ask for a timeline for AGI, 76% of respondents shared that scaling up current AI approaches would be unlikely to lead to AGI.30

OpenAI expands its robotics ambitions

OpenAI is increasing its focus on robotics as part of its goal to advance artificial general intelligence. The company is hiring specialists in humanoid robot systems and forming a team to design algorithms that help robots learn and act independently in the physical world.

This marks a shift from OpenAI’s earlier focus on language and image models. The company now aims to connect advanced reasoning with physical interaction, suggesting it views robotics as an essential step toward testing and achieving AGI.

Context and implications

After winding down its first robotics team around 2020, OpenAI is returning to active development in the field. Recent hiring and potential partnerships point to a renewed effort to build robots capable of real-world learning and manipulation.

By combining large-scale AI models with sensory data, OpenAI aims to create systems that can reason and operate outside digital environments. The recruitment of humanoid robotics experts also indicates long-term goals that go beyond automation and toward robots that can work safely alongside people.31

Microsoft’s report on early experiments with GPT-4

Microsoft Research studied an early version of OpenAI’s GPT-4 in 2023. The report claimed that it showed greater general intelligence than previous AI models, performing at a human level in areas like math, coding, and law. This sparked debate on whether GPT-4 was a preliminary form of artificial general intelligence. 32

The road to artificial general intelligence report by MIT

“The road to artificial general intelligence” report in August 2025 anticipates that early AGI-like systems could begin emerging between 2026 and 2028, showing human-level reasoning within specific domains, multimodal capabilities across text, audio, and physical interfaces, and limited goal-directed autonomy.

The report combines aggregated forecasts and suggests a 50% probability that several generalized milestones, such as knowledge transfer and broad reasoning, will be achieved by 2028.

Longer-range projections estimate that machines may surpass human performance in all economically valuable tasks by around 2047, contingent on advances in compute efficiency, algorithmic breakthroughs, and autonomous learning.33

AI Frontiers on AGI probabilities

Adam Khoja and Laura Hiscott from AI Frontiers, a platform for AI debates and dialogues, estimate a 50% probability of reaching AGI by 2028 and an 80% probability by 2030, using their quantitative AGI definition.34

Khoja and Hiscott evaluate progress toward artificial general intelligence using a definition developed by Khoja, Dan Hendrycks, and their co-authors.35 Their framework measures ten cognitive abilities and assigns GPT-4 a score of 27% and GPT-5 a score of 57%. This indicates that current models are roughly halfway to the defined AGI threshold.

Khoja and Hiscott argue that traditional discussions about AGI timelines lack precision because they rely on inconsistent definitions. Their standardized framework is intended to create clarity by identifying specific strengths and weaknesses in current models. They note that reading, writing, mathematics, and general knowledge meet or exceed human baselines and are no longer limiting factors.

The authors highlight remaining gaps in visual reasoning, intuitive physics, auditory processing, perception-dependent speed, and visual and auditory working memory. They report rapid improvement on benchmarks such as SPACE and MindCube and suggest these gaps can likely be addressed through continued incremental research. They also observe that hallucinations remain a concern but are tractable given performance differences across leading models.

According to Khoja, Hiscott, and Hendrycks, the most significant remaining obstacle is continual learning and long-term memory storage. Current systems cannot retain information across sessions, and resolving this limitation will require at least one meaningful breakthrough. However, the authors emphasize that major AI labs are now prioritizing this area.

Learning from past over-optimism in AI predictions

Keep in mind that AI researchers were over-optimistic before. Examples include:

This historical experience contributed to most current scientists shying away from predicting AGI in bold time frames like 10-20 years, but this has changed with the rise of generative AI.

Understand what singularity is

Artificial intelligence scares and intrigues us. Almost every week, there’s a new AI scare on the news, like developers afraid of what they’ve created or shutting down bots because they got too intelligent.39

Most of these myths result from research misinterpreted by those outside the AI and GenAI fields. Some stakeholders claim to fear AI because they may profit from more regulation, or it may bring them more attention.

The greatest fear about AI is singularity (also called Artificial General Intelligence or AGI), which is an event that is expected to bring a rapid increase in machine intelligence. This is expected when a system combines human-level thinking with superhuman speed and rapidly accessible, near-perfect memory. According to some experts, singularity also implies machine consciousness.

Such a machine could self-improve and surpass human capabilities. Even before artificial intelligence was a computer science research topic, science fiction writers like Asimov were concerned about this. They were devising mechanisms (i.e., Asimov’s Laws of Robotics) to ensure the benevolence of intelligent machines, which is more commonly called alignment research today.

Why experts believe AGI is inevitable: Key arguments & evidence

Reaching AGI feels like a wild prediction, but it seems like quite a reasonable goal when you consider that human intelligence is fixed and machine intelligence is growing. It is only a matter of time before machines surpass us unless there’s some hard limit to their intelligence. We haven’t encountered such a limit yet.

Human intelligence is fixed unless we somehow merge our cognitive capabilities with machines. Elon Musk’s neural lace startup aims to do this, but research on brain-computer interfaces is in the early stages.40

Machine intelligence depends on algorithms, processing power, and data.

Recent achievements

Opus 4.6

On February 2026, Claude released Opus 4.6 with a 1M context window and impressive benchmark results.

Anthropic is also focusing on use cases by releasing plugins like Claude legal which are markdown files for helping models navigate specific domains. Though this was a minor addition to Claude, it triggered a stock market selloff including SaaS and legal software.41

Gemini Deep Think

Another example is DeepMind’s Gemini deep think mode, which achieved gold-medal performance at the 2025 International Mathematical Olympiad, marking a significant step in AI’s ability to reason through complex problems.

Operating entirely in natural language, Gemini solved five out of six problems within the official 4.5-hour contest window, while producing clear, human-readable proofs without relying on formal symbolic tools.

Its capabilities stem from several innovations: Deep Think mode enables parallel exploration of solution paths, training incorporates expert-level mathematical proofs, and reinforcement learning refines its strategic approach.

This progress demonstrates that advanced AI can now engage in sophisticated, interpretable reasoning at a level once reserved for top human problem-solvers.42

Opencrawl

Opencrawl is an open source project to turn LLMs into agents. It became one of the most popular projects on GitHub and kicked off the opencrawl ecosystem.

Exponential growth

The following is a helpful analogy for understanding exponential growth. While machines may not seem highly intelligent right now, they can become quite smart in the near future.

Recent growth in AI computing capabilities

Figure 2: The figure shows a summary of the compute growth patterns observed across various categories: overall notable models (top left), frontier models (top right), leading language models (bottom left), and top models from leading companies (bottom right).

Computational resources for training AI models have significantly increased, with about two-thirds of language model performance attributed to model scale improvements.

According to a 2024 article,43 the growth of compute usage in training AI models has consistently increased by around 4-5x per year, reflecting trends in notable models, frontier models, and top companies like OpenAI, Google DeepMind, and Meta AI (See Figure 2).

However, the growth rate has slowed somewhat since 2018, especially for frontier models, but language models have experienced faster growth up to 9x/year until mid-2020, after which the pace slowed to 4-5x/year.

The overall trend for AI compute growth remains strong, and projections suggest that the growth rate of 4-5x/year will continue unless new challenges or breakthroughs occur. This growth is also seen in the scaling strategies of leading AI companies, though slight variations exist between them.

Despite a slowdown in frontier model growth, the larger models released today, such as GPT-4 and Gemini Ultra, align closely with the predicted growth trajectory.

If classic computing slows, quantum computing may fill the gap

Classic computing has taken us quite far. AI algorithms on classical computers can exceed human performance in specific tasks like playing chess or Go. For example, AlphaGo Zero beat AlphaGo by 100-0. AlphaGo had beaten the best players on earth.44 However, we are approaching the limits of how fast classical computers can be.

Moore’s law, which is based on the observation that the number of transistors in a dense integrated circuit doubles about every two years, implies that the cost of computing halves approximately every 2 years.

On the other hand, most experts believe that Moore’s law is coming to an end during this decade.45 However, there are efforts to keep improving the efficiency of computing.

For example, DeepSeek surprised global markets with its R1 model by delivering a reasoning model at a fraction of the cost of its competitors, like OpenAI.

Quantum Computing, which is still an emerging technology, can contribute to reducing computing costs after Moore’s law comes to an end. Quantum Computing is based on the evaluation of different states at the same time, whereas classical computers can calculate one state at a time.

The unique nature of quantum computing can be used to efficiently train neural networks, currently the most popular AI architecture in commercial applications. AI algorithms running on stable quantum computers have a chance to unlock the singularity.

Why do some experts believe that we will not reach AGI?

There are 3 major arguments against the importance or existence of AGI. We examined them along with their common rebuttals:

1- Intelligence is multi-dimensional

Therefore, AGI will be different, not necessarily superior to human intelligence.

This is true, and human intelligence is also different from animal intelligence. Some animals are capable of mental feats, like squirrels remembering where they hid hundreds of nuts for months.

Yann LeCun, one of the pioneers of deep learning, believes that we should retire the word AGI and focus on achieving “advanced machine intelligence”.46 He argues the human mind is specialized and intelligence is a collection of skills and the ability to learn new skills. Each human can only accomplish a subset of human intelligence tasks.47

It is also hard to understand the specialization level of the human mind, as humans, since we don’t know and can’t experience the entire spectrum of intelligence.

In areas where machines exhibited super-human intelligence, humans were able to beat them by leveraging machine-specific weaknesses. For example, an amateur was able to beat a Go program that is on par with Go programs that beat world champions by studying and leveraging the program’s weaknesses.48

2- Intelligence is not the solution to all problems

Science

Even the best machine analyzing existing data may not be able to find a cure for cancer. It may need to run real-world experiments and analyze results to discover new knowledge in most areas.

More intelligence can lead to better-designed and managed experiments, enabling more discovery per experiment. The history of research productivity should demonstrate this, but the data is quite noisy, and there are diminishing returns on research. We encounter harder problems like quantum physics as we solve simpler problems like Newtonian motion.

Finally, perfect predictions may not be possible in some domains due to the inherent randomness or immeasurability of that domain. For example, even with a wealth of data, we are not able to predict certain life outcomes with a high level of accuracy.49

Economy

Intelligence is not the only ingredient for economic value generation.

Figure 3: IQ is correlated with wealth at low levels of wealth.50

Figure 4: IQ is not correlated with wealth if we only focus on high levels of wealth. This graph is the same as the one above except that net income levels below $40k have been hidden51

3- AGI is not possible because it is not possible to model the human brain

Theoretically, it is possible to model any computational machine, including the human brain, with a relatively simple machine that can perform basic computations and access infinite memory and time. This is the universally accepted Church-Turing hypothesis laid out in 1950. However, as stated, it requires certain difficult conditions: infinite time and memory.

Most computer scientists believe that modeling the human brain will take less than infinite time and memory. Nonetheless, there is no mathematically sound way to prove this belief, because we do not yet understand the brain well enough to precisely characterize its computational power. We will have to build such a machine!

How can we reach AGI?

Figure 5: The time horizon of frontier AI models over time shows the longest tasks (in human-equivalent time) each model can complete with 50% reliability.52

The above figure shows how AI agents‘ capabilities have improved over time by measuring the longest tasks they can complete with 50% reliability.

The key finding is that the task length frontier models can handle has grown exponentially, doubling roughly every seven months. This means newer models, like Claude 3.7 Sonnet and o1, can now complete tasks that would take a human nearly an hour, while older models like GPT-2 could barely handle tasks longer than a few seconds.

The shaded region reflects statistical uncertainty, but the overall trend is reliable. If this pattern continues, AI systems could soon handle complex tasks that take humans days or even weeks, marking a significant step toward broader autonomy and AGI-like capabilities.

Scaling as a pathway to AGI

Leaders of frontier AI labs believe that scaling current transformer-based approaches can yield AGI, which fuels their predictions about achieving AGI in a few years.

One proposed pathway to AGI is scaling up existing architectures like transformers by increasing compute and data, while another is developing entirely new approaches.

In support of the scaling hypothesis, a 2024 report by Epoch AI analyzed whether AI compute growth can continue through 2030.

They identified four major constraints: power availability, chip manufacturing capacity, data scarcity, and processing latency (See Figure 6).

Despite these challenges, they argue it’s feasible to train models requiring up to 2e29 FLOPs by the end of the decade, assuming significant investments in infrastructure.

Such advancements could produce AI systems far more capable than today’s state-of-the-art models like GPT-4, pushing us closer to AGI.53

Figure 6: The chart illustrates the estimated upper bounds on AI training compute by 2030 under key constraints, power, chip production, data, and latency, with medians ranging from 2e29 to 3e31 FLOP.

Beyond scaling: The case for new architectures

However, influential AI scientists like Yann LeCun and Richard Sutton believe that scaling large language models will not lead to human-level intelligence.54 55 They believe that new architectures or approaches are necessary for AGI.

How can we measure whether we have reached AGI?

Large language models are blowing past new benchmarks every week, but evaluating LLMs is difficult due to issues like data poisoning and the lack of an accepted scientific definition for human-level intelligence.

These concerns are amplified by insights from recent research56 which highlight that scaling LLMs is not a sustainable path to better performance, especially in scientific and high-stakes domains. The authors show that:

These findings call into question the reliability of standard benchmarks and underscore the need for more diverse and evolving evaluation strategies.

Old metrics like the Turing test are no match for today’s machines, and new metrics like ARC-AGI may lack the generalization capabilities of broader benchmarks.

Emerging metrics like ARC-AGI aim to test abstraction and generalization, but may still lack resilience to data contamination or overfitting.

Moreover, as the paper highlights, even “good” loss scores may mask underlying information catastrophes due to non-Gaussian fluctuations and training instabilities.57

How can we track the progress of LLMs?

There are a few approaches to benchmarking to overcome these challenges:

What are approaches beyond benchmarking to determine AGI?

There are potentially strong but lagging indicators of the impact of AI, which can help identify AGI.

Economic growth

Microsoft CEO Satya Nadella claims that 10% growth in the developed world would indicate AGI.58 . However, his incentive is to have a delayed definition of AGI since AGI would end OpenAI and Microsoft’s exclusive partnership.59

Unemployment

We expect AGI to

In a world where machines are more intelligent and efficient than humans, it wouldn’t be rational to pay a human to sit in front of a computer. Therefore, we expect white collar employment to plummet while humans continue to thrive in jobs in the physical world.

Government agencies collecting labor statistics classify jobs into detailed categories, making white collar employment an easy-to-track metric.

We gathered data from the U.S. Bureau of Labor Statistics on white collar employment spanning 2019 to 2024.60 For clarity and consistency, we categorized white collar workers into the following occupational groups:

According to our analysis, the ratio of white collar workers to total employment has fluctuated between 45% and 48% over this period.

While this range suggests relative stability in the share of white collar employment so far, it is not indicative of a long-term trend, and we expect more pronounced shifts in the coming years as automation and AI adoption accelerate. For more predictions on how AI will change white collar and entry-level employment, read AI job loss.

Shall we even aim for AGI?

There are computer scientists who warn that focusing on AGI as the ultimate goal may distort AI research.61 Criticisms include: Creating an illusion of consensus, overfitting benchmarks, ignoring embedded social values, letting hype dictate priorities, building up “generality debt” (postponing key design questions), and excluding marginalized communities and under-resourced researchers.

Specific, measurable, and transparent goals would be better for progress in AI than a vaguely defined goal like AGI.

Mathematical reasoning behind AGI predictions

Mathematical reasoning is central to understanding and forecasting AGI timelines. Many projections are based on quantifiable trends and formal models that guide expectations about when artificial general intelligence might emerge.

Scaling laws and compute growth

One key component of mathematical reasoning involves analyzing scaling laws. These show that model performance improves predictably with more data, parameters, and compute.

The consistent 4–5× annual growth in AI training compute supports forecasts that AGI may be achievable within one or two decades, assuming current trends continue.

These projections are based on empirical fits to performance curves and extrapolations, underpinned by power-law relationships, a core concept in mathematical modeling.

Probabilistic forecasting

Researchers also apply probabilistic methods to AGI predictions. Surveys often ask experts to estimate the probability of AGI being developed by specific years, producing cumulative probability distributions.

For example, a 50% probability by 2040 reflects consensus under uncertainty, driven by Bayesian-style updating based on observed AI progress.

This mathematical reasoning approach captures expert uncertainty without requiring precise dates, allowing ongoing revision as new data becomes available.

Theoretical foundations

These forecasts are based on theoretical elements of mathematical reasoning, including the Church-Turing thesis, which implies that human cognition can be simulated by machines, and concepts like Kolmogorov complexity, which relate intelligence to the compressibility of information.

While such theories do not guarantee AGI, they provide a framework for thinking about its possibility and the computational requirements involved.

More about Artificial General Intelligence

David Silver, Principal Research Scientist at Google DeepMind

He explains that Artificial General Intelligence (AGI) refers to AI systems capable of learning and excelling at a wide range of tasks; much like humans who can become experts in diverse fields such as science, music, or sports.

Unlike narrow AI limited to a single function, AGI aspires to mirror human adaptability and general problem-solving ability.

He notes that while AGI is a long-term goal, reaching true human-level intelligence will likely require several breakthroughs and will develop gradually over time (See the video below).

David Silver of DeepMind describes AGI as AI with human-like versatility across tasks, noting it will require multiple breakthroughs and will develop gradually over time.

Ilya Sutskever, co-founder and Chief Scientist of OpenAI

In the TED Talk “The Exciting, Perilous Journey Toward AGI,” he explores the rapid progress toward Artificial General Intelligence (AGI).

He predicts AGI could emerge within the next 5 to 10 years, though he acknowledges uncertainty in this timeline.

Sutskever highlights both the immense potential and the profound risks of AGI, stressing the need to align its development with human values. Despite the challenges, he is optimistic that humanity can safely guide this powerful technology (See the video below).

In his TED Talk, Ilya Sutskever predicts AGI could arrive within 5–10 years, emphasizing its transformative potential and the urgent need to align it with human values to ensure a safe future.

Ray Kurzweil, computer scientist and entrepreneur

He reflects on over six decades of AI progress, tracing humanity’s ability to build intelligence-enhancing tools, from primitive implements to large language models.

He also predicts that Artificial General Intelligence will arrive by 2029, leading to technological singularity by 2045. He highlights exponential advances in computing power, medicine, and biotechnology.

He also forecasts breakthroughs like AI-generated cures, digital clinical trials, and longevity escape velocity, where scientific progress could extend life indefinitely (See the video below).

In his TED Talk, Ray Kurzweil predicts AGI by 2029 and a technological singularity by 2045, envisioning a future where exponential AI advances revolutionize medicine and extend human longevity.

Yann LeCun, Turing award recipient

See why LLMs can not give us human-level intelligence and the latest AI approaches to get there:

Artificial general intelligence predictions

Singularity graph methodology

To plot the expected year of AGI development on the graph, we used the weighted average of predictions for each year within each category. For example, if there were multiple Prediction Market forecasts in 2022, we calculated their weighted average and plotted that value.

Conclusion

Predictions for AGI have shifted notably in recent years. While earlier surveys placed its arrival closer to 2060, recent forecasts, especially from entrepreneurs, suggest it could emerge as early as 2026–2035.

This change is fueled by rapid advances in large language models and growing compute power. Yet, despite these gains, today’s AI still lacks the general flexibility and autonomy associated with human-level intelligence.

Experts remain divided on how AGI will be achieved; some believe scaling current architectures will be enough, while others argue that new methods are needed.

Key challenges include high resource demands, unclear benchmarks, and unresolved ethical concerns. AGI may be closer than ever, but its arrival still hinges on both technical breakthroughs and careful oversight.

FAQs

Singularity is a hypothetical event which is expected to result in a rapid increase in machine intelligence.

For singularity, we need a system that combines human-level thinking with superhuman speed and rapidly accessible, near-perfect memory.

Singularity should also result in machine consciousness but since consciousness is not well-defined, we can’t be precise about it. Such a system could self-improve and surpass human capabilities.

While singularity is a relatively old term, AGI and especially superintelligence are used more frequently these days to describe the same event.

Artificial General Intelligence (AGI) refers to a type of AI that can understand, learn, and apply knowledge across a broad range of intellectual tasks at a level equal to or exceeding that of humans.

Unlike narrow AI, which is designed for specific tasks such as language translation or image recognition, AGI would possess generalized cognitive abilities, enabling it to reason, plan, and adapt in unfamiliar situations.

The development of AGI remains a significant research goal and subject of ethical and philosophical debate.

Superintelligence denotes an intellect that significantly surpasses the best human minds in virtually all domains, including creativity, problem-solving, and social understanding.

It represents a stage beyond AGI, where an artificial system could outperform humans in every meaningful intellectual pursuit.

The concept raises critical considerations about control, safety, and the long-term implications for humanity’s role in a world dominated by superior intelligence.

Advanced Machine Intelligence (AMI) involves competent AI systems that approach or achieve near-general intelligence.

While they may not yet possess the complete flexibility and self-awareness associated with AGI, AMI systems demonstrate advanced reasoning, learning, and adaptability across diverse tasks.

The term is often used to denote AI systems that exceed current narrow AI capabilities but remain below the threshold of complete general intelligence.

Cite this benchmark

Pick the format that matches where you're publishing. Pasting the link version into your CMS preserves the backlink.

Cem Dilmegani and Sıla Ermut (2026) - "AGI/Singularity: 9,800 Predictions Analyzed". Published online at AIMultiple.com. Retrieved May 22, 2026, from: https://aimultiple.com/artificial-general-intelligence-singularity-timing [Online Resource]

Dilmegani, C., & Ermut, S. (2026, May 22). AGI/Singularity: 9,800 Predictions Analyzed. AIMultiple. https://aimultiple.com/artificial-general-intelligence-singularity-timing

@misc{dilmegani2026, author = {Dilmegani, Cem and Ermut, Sıla}, title = {{AGI/Singularity: 9,800 Predictions Analyzed}}, year = {2026}, month = may, howpublished = {\url{https://aimultiple.com/artificial-general-intelligence-singularity-timing}}, note = {AIMultiple. Retrieved May 22, 2026} }

Results and timestamps of 204 data points. Download the data used in this article as a ZIP file containing 2 CSV files and a README.

Cem Dilmegani

Cem Dilmegani

Principal Analyst

Cem has been the principal analyst at AIMultiple since 2017. AIMultiple informs hundreds of thousands of businesses (as per similarWeb) including 55% of Fortune 500 every month.

Cem's work has been cited by leading global publications including Business Insider, Forbes, Washington Post, global firms like Deloitte, HPE and NGOs like World Economic Forum and supranational organizations like European Commission. You can see more reputable companies and resources that referenced AIMultiple.

Throughout his career, Cem served as a tech consultant, tech buyer and tech entrepreneur. He advised enterprises on their technology decisions at McKinsey & Company and Altman Solon for more than a decade. He also published a McKinsey report on digitalization.

He led technology strategy and procurement of a telco while reporting to the CEO. He has also led commercial growth of deep tech company Hypatos that reached a 7 digit annual recurring revenue and a 9 digit valuation from 0 within 2 years. Cem's work in Hypatos was covered by leading technology publications like TechCrunch and Business Insider.

Cem regularly speaks at international technology conferences. He graduated from Bogazici University as a computer engineer and holds an MBA from Columbia Business School.

View Full Profile

Researched by

Sıla Ermut

Sıla Ermut

Industry Analyst

Sıla Ermut is an industry analyst at AIMultiple focused on email marketing and sales videos. She previously worked as a recruiter in project management and consulting firms. Sıla holds a Master of Science degree in Social Psychology and a Bachelor of Arts degree in International Relations.

View Full Profile