The Global Consciousness Project Explained: RNGs, 9/11 and Doubts

The Global Consciousness Project (GCP) is a long-running experiment, started in 1998 by psychologist Roger D. Nelson, that runs hardware random number generators at sites around the world to test whether moments of shared human attention or emotion line up with non-random structure in their output. Its organizers report a cumulative deviation of more than seven standard deviations across about 500 pre-registered events, while critics and some of its own former analysts argue the result reflects flexible analysis choices rather than a global mind.

The project matters because it is the most sustained attempt to put a number on an old idea: that consciousness is not sealed inside individual skulls but is in some sense shared. That idea sits at the center of many spiritual traditions and of Billy Carson's own teaching. The GCP took it out of philosophy and into a database, which means it can be checked, and the checking is where the real story lies.

Where did the Global Consciousness Project come from?

The GCP grew out of the Princeton Engineering Anomalies Research (PEAR) laboratory. PEAR was founded in 1979 by Robert G. Jahn, then dean of Princeton's School of Engineering and Applied Science and an authority on spacecraft propulsion, with developmental psychologist Brenda Dunne as laboratory manager. Its best-known experiment asked volunteers to sit in front of an electronic random event generator (REG), a device producing a stream of random binary outputs, and try by intention alone to push the output higher or lower than chance. The lab also ran remote perception studies, the same family of experiments discussed in our guide to remote viewing and the Stargate program.

Over millions of trials PEAR reported a very small effect, on the order of a few parts in 10,000, that became statistically significant because of the enormous number of trials. Replication was the sticking point. A three-laboratory consortium with research groups in Freiburg and Giessen, Germany, published in the Journal of Scientific Exploration in 2000, found that its primary measure did not replicate the original result, although all three labs saw small positive trends.

Jahn and Dunne closed the lab at the end of February 2007, after roughly 28 years, as Jahn prepared to retire. He told The New York Times that if people did not believe the lab's results by then, they never would. Physicist Robert L. Park of the University of Maryland called the work an embarrassment to science and to Princeton, while others defended Jahn's freedom to pursue unpopular questions. Jahn died in 2017.

Roger Nelson was PEAR's coordinator of experimental work from 1980 until he retired in 2002. In 1997, field studies that took portable generators into group events led him to conceive a permanent worldwide network. The GCP is independent of PEAR and of Princeton University; it was funded by private donations made through the Institute of Noetic Sciences.

How does the RNG network actually work?

The network is a set of hardware random number generators, each attached to a computer at a host site, that report their output to a central server every second. Data collection began in August 1998 with three devices. The network grew to around 30 by about 2000 and plateaued at roughly 65 to 70 reporting devices around 2004, spread across several continents.

Each device produces a "trial" once per second: the sum of 200 random bits. If the bits are truly random, each trial averages 100 with a standard deviation of about 7.07, and the long-run output of every device should look like a coin being flipped fairly. The devices became known as "eggs", a play on "electrogaiagram" (EGG), a blend of electroencephalogram and Gaia, so the effort was first called the EGG project.

The working hypothesis is simple to state. When large numbers of people focus on the same thing at the same time, such as New Year's midnight, a major disaster or a coordinated mass meditation, the outputs of the separate devices should become slightly more correlated with one another or drift slightly from expectation. The GCP describes this as structure in the data that should not be there if the devices are independent and random.

How were events tested and scored?

The GCP's central safeguard was a prediction registry. For each formal test, the event, the start and end of the time window, and the statistic to be applied were supposed to be written down before anyone looked at the data. Each event then produced a single Z-score, a measure of how far that window departed from chance, and the scores from all events were combined into a running total.

The formal series ran from 1998 to December 2015 and included about 500 registered events. Nelson reports that the combined result departs from chance expectation by more than seven sigma, which he expresses as odds of roughly a trillion to one against chance.

This is the GCP's strongest point and its most contested one. A pre-registered, cumulative design is good practice, and the project published its registry and raw data, which allowed outsiders to reanalyze it. But the registry left the researchers free to choose which events counted as "global," how long each window should be and which statistic to use, and critics argue those choices carry most of the weight.

What did the GCP report on September 11, 2001?

The GCP's most cited event is 11 September 2001. Its analysts reported a strong deviation from randomness over a window of roughly four hours that began shortly before the first plane struck, and in exploratory analysis they described continuing deviations over the following days. The project presented the day as a clear example of its hypothesis.

Independent researchers Edwin May and James Spottiswoode reanalyzed the same data in 2002. They found that the significance depended on the exact window chosen: shortening it by a few minutes or extending it by about half an hour produced results consistent with chance. Nelson responded that their approach was itself an unacceptable form of post hoc data selection, because the registered window, not alternative windows chosen later, was the proper test.

NASA astrophysicist Jeffrey Scargle, writing in 2002, took a broader line. He argued that a claim of this kind would not be generally accepted unless both classical significance testing and Bayesian analysis agreed and showed the same anomalous effect. On 9/11 and similar events, that agreement has not been demonstrated. The devices did not "predict" the attacks; the reported deviation is a statistical pattern in a time window, not a warning.

What do skeptics and later analysts say?

The mainstream position is that the GCP's result is unconfirmed and that several well-documented methodological objections stand against it. Robert Todd Carroll of The Skeptic's Dictionary and Claus Larsen were among the early critics. Their main points were:

• Flexible choices. Deciding which events qualify as global and how long their windows run leaves room for unconscious bias, even with a registry.
• Pattern-matching and selection. With hundreds of events and many ways to slice the data, some significant-looking results are expected by chance.
• No mechanism. The project has not proposed a physical process by which shared attention would alter electronic noise.
• No independent replication. No separate group has built its own network and reproduced the effect.

The most significant critique came from inside. Peter Bancel, formerly one of the GCP's analysts, published "Searching for global consciousness: A 17-year exploration" in the journal Explore in 2017. He concluded that the data do not support the global-consciousness proposal and that the pattern of results fits better with a goal-oriented or experimenter effect, in which the outcome tracks the researchers' expectations rather than the size of the world's attention. Nelson published a response defending the original interpretation.

The wider literature points the same way. A 2006 meta-analysis in Psychological Bulletin by Holger Bösch, Fiona Steinkamp and Emil Boller pooled hundreds of studies in which people tried to influence random number generators by intention. It found a tiny overall effect, larger effects in smaller studies, and evidence that publication bias could in principle account for the result. Dean Radin, Nelson and colleagues disputed that conclusion in a published reply, and the debate over how to read such small effects continues.

What is happening with the project now?

According to the Society for Psychical Research's Psi Encyclopedia, the original network kept collecting data and offering public access until April 2026, when its hosting servers suffered a catastrophic failure; the encyclopedia reports that the archival data and a restored website are to be maintained.

A successor, GCP 2.0, is run by the HeartMath Institute with research operations directed by Rollin McCraty, working with Nelson, the Institute of Noetic Sciences and Dean Radin as chief scientist. It was planned in the early 2020s and launched in 2023 and 2024. The design calls for about 1,000 new devices, each containing four independent generators, with many placed in city clusters. McCraty and colleagues published preliminary results in 2024 in the Journal of Management, Spirituality & Religion. Those results have not been independently replicated, and the project states that it intends to keep a public hypothesis registry and give outside researchers access to raw data.

Billy Carson's reading

No first-party statement by Billy Carson on the Global Consciousness Project has been found, so this section does not put words in his mouth. What his work does document is a clear view of mind. In his teaching on the universal mind, Carson holds that "All that exists is the Universal Mind and nothing more than that one mind," and that "There is only one intelligence, one mind in all creation, and we are all part of it."

That is precisely the idea the GCP set out to test. The question Carson's framework raises is direct: if consciousness is shared, should it leave a measurable trace in physical randomness? The GCP is the best-known attempt to answer that question with instruments, and the honest summary is that it produced a striking cumulative number and an unresolved argument about what the number means. A larger network with open data, such as GCP 2.0, is the kind of test that could strengthen or weaken the case.

Carson's own emphasis is practical rather than statistical: training attention and intuition. His 4BK TV series Awaken the 6th Sense walks through psychic-ability and remote-viewing exercises, and the same theme runs through our guide to synchronicity and Jung's meaningful coincidence, which deals with the inner experience of connection that the GCP tried to measure from the outside.

Watch, read and go deeper

Start with Awaken the 6th Sense on 4BK TV for Billy Carson's exercises in intuition and remote viewing. For the research side, read our explainer on the science of ESP and the guide to what meditation research shows, since mass meditations are among the events the GCP registered. The consciousness topic hub gathers every article, episode and talk on the subject in one place.

Frequently asked questions

Did the Global Consciousness Project prove that a global consciousness exists?

No. The GCP reports a cumulative deviation of more than seven sigma across about 500 registered events, but that is a statistical anomaly, not proof of a shared mind. Independent reanalyses, critiques of its event and window choices, and a former analyst's 2017 review argue that the result may come from analysis choices or experimenter effects. The mainstream view is that the claim is unconfirmed.

Was the Global Consciousness Project run by Princeton University?

No. It grew out of the Princeton Engineering Anomalies Research lab, where founder Roger Nelson worked from 1980 to 2002. The GCP itself was an independent project funded by private donations through the Institute of Noetic Sciences.

What are the "eggs" in the Global Consciousness Project?

The eggs are hardware random number generators placed at host sites around the world. Each one records the sum of 200 random bits every second and sends it to a central server. The nickname comes from "electrogaiagram," a blend of electroencephalogram and Gaia.

Why is the 9/11 result disputed?

The GCP reported a strong deviation in a window of about four hours around the attacks. Edwin May and James Spottiswoode found that shifting that window by a few minutes or about half an hour made the result consistent with chance, suggesting the outcome depended on the window chosen. Nelson rejected their reanalysis as post hoc.

Is the Global Consciousness Project still running?

The original network's data collection ended in April 2026 when its hosting servers failed, according to the Psi Encyclopedia, which reports that archival data is being preserved. A successor, GCP 2.0, run by the HeartMath Institute with Nelson and Dean Radin, is building a larger network of about 1,000 four-generator devices. Its preliminary results, published in 2024, have not been independently replicated.

Sources and further reading

• Roger D. Nelson, "Global Consciousness Project," Psi Encyclopedia, Society for Psychical Research.
• Roger D. Nelson, "Princeton Engineering Anomalies Research (PEAR) Laboratory," Psi Encyclopedia, Society for Psychical Research.
• Robert G. Jahn et al., "Mind/Machine Interaction Consortium: PortREG Replication Experiments," Journal of Scientific Exploration 14(4), 2000.
• Jeffrey D. Scargle, "Was There Evidence of Global Consciousness on September 11, 2001?" Journal of Scientific Exploration, 2002.
• Peter A. Bancel, "Searching for Global Consciousness: A 17-Year Exploration," Explore, 2017.
• Holger Bösch, Fiona Steinkamp and Emil Boller, "Examining Psychokinesis: The Interaction of Human Intention With Random Number Generators," Psychological Bulletin, 2006.
• Robert Todd Carroll, The Skeptic's Dictionary, entries on the Global Consciousness Project and PEAR.
• Rollin McCraty et al., preliminary GCP 2.0 results, Journal of Management, Spirituality & Religion, 2024.