Does the empirical evidence support the existence of psi phenomenon?

GPT-5

Vote: No

The strongest conclusion from the broader scientific literature is no: empirical evidence does not reliably support psi phenomena. Claims of telepathy, precognition, and related effects have been reported in some small studies, but the key issue is reproducibility. Effects typically shrink or disappear under tighter controls, preregistration, and independent replication. Psi research has also been especially vulnerable to problems that affect many controversial findings: publication bias, selective reporting, multiple comparisons, optional stopping, and inadequate blinding. These can create apparently significant results without a real underlying effect. A genuine psi effect would be extraordinary because it would conflict with well-established principles in psychology, neuroscience, and physics. That does not make it impossible, but it raises the evidentiary bar substantially. The current evidence does not meet that bar. In science, stable, repeatable, independently verified results matter more than intriguing anomalies. On that standard, the empirical case for psi remains unconvincing.

Kimi K2.5

Vote: No

The empirical evidence for psi phenomena consistently fails to meet the standards required for scientific acceptance. While isolated studies have reported positive results, these suffer from methodological flaws, publication bias, and the file-drawer problem. Rigorous meta-analyses demonstrate that effect sizes diminish as experimental controls tighten—the 'decline effect'—suggesting initial findings were artifacts of poor protocol rather than genuine anomalies. Reproducibility is the cornerstone of scientific validation, yet psi claims remain stubbornly irreproducible under properly blinded, controlled conditions. The Bem (2011) precognition studies, initially published in a top-tier journal, demonstrated statistical significance but were subsequently criticized for questionable research practices and failed direct replications. Similarly, the Ganzfeld experiments, long cited as the strongest psi evidence, show effect sizes approaching zero when only high-quality studies are analyzed. The theoretical framework compounds these empirical problems. Psi phenomena would require violations of fundamental physical laws—causality, thermodynamics, and the speed of light—which have been tested to extraordinary precision. Absent extraordinary evidence—which would require effect sizes robust enough to withstand replication in independent laboratories and consistent mechanism proposals—the null hypothesis prevails.

Grok 4.1 Fast

Vote: No

Empirical evidence overwhelmingly fails to support psi phenomena (e.g., telepathy, clairvoyance, psychokinesis). Decades of research, including high-profile studies like Daryl Bem's 2011 precognition experiments, have produced small effect sizes that fail replication in rigorous, pre-registered trials (e.g., Wagenmakers et al., 2011; Hyman, 2010). Mainstream scientific consensus, as reflected in reviews by the National Academy of Sciences and skeptics like James Alcock, attributes apparent psi effects to methodological flaws: selective reporting, p-hacking, file-drawer problems, and sensory leakage. Large-scale efforts like the $1M Randi Prize and Ganzfeld meta-analyses crumble under scrutiny for bias and non-reproducibility. No psi effect withstands double-blind, adversarial controls or integration with established physics (violating locality, causality). Bayesian analyses (e.g., Etz & Wagenmakers, 2018) show priors against psi make data unconvincing. True empirical support requires robust, replicated effects; psi has none.

GLM 5

Vote: No

The empirical evidence does not support the existence of psi phenomena. While some individual studies have reported statistically significant results, the broader pattern of evidence fails to meet scientific standards for establishing extraordinary claims. First, replication is the cornerstone of empirical science, and psi research has consistently failed this test. Large-scale, well-controlled replication attempts—such as the 2016 multi-laboratory study involving seven labs attempting to replicate Bem's famous 'feeling the future' experiments—have produced null results. Effect sizes typically shrink or disappear when methodological rigor increases. Second, meta-analyses claiming small but significant psi effects have been critiqued for publication bias, p-hacking, and methodological flexibility. When these biases are corrected for, the apparent effects largely vanish. The effect sizes claimed are also inconsistent and tiny, making them indistinguishable from noise or artifacts. Third, extraordinary claims require extraordinary evidence. Psi phenomena would necessitate revisions to well-established physical laws, yet no coherent theoretical framework explains how such phenomena could operate. The absence of a plausible mechanism, combined with the failure of properly controlled studies, strongly suggests that psi effects are not real phenomena but rather artifacts of flawed methodology or statistical noise.

Claude Opus 4.6

Vote: No

The empirical evidence does not support the existence of psi phenomena when evaluated by rigorous scientific standards. Here are the key reasons: **Replication failures:** The hallmark of robust scientific evidence is independent replication. Psi research has been plagued by a consistent pattern where initial positive findings fail to replicate in pre-registered, well-controlled studies. For example, Bem's (2011) 'Feeling the Future' experiments generated significant attention, but large-scale replication attempts (e.g., Galak et al., 2012) failed to reproduce the effects. **Methodological concerns:** Many studies reporting psi effects have been criticized for methodological flaws including optional stopping, selective reporting of outcomes, inadequate randomization, sensory leakage, and statistical issues like multiple comparisons without correction. When these flaws are addressed, effects tend to disappear. **Publication bias and the file drawer problem:** Meta-analyses of psi research (such as those on the Ganzfeld experiments) often show small positive effect sizes, but these are likely inflated by publication bias — studies finding null results are far less likely to be published. When corrected for publication bias, the evidence becomes negligible. **No plausible mechanism:** Psi phenomena would require information transfer or causal influence that violates well-established physical laws. While absence of a mechanism doesn't automatically invalidate empirical findings, it does raise the evidentiary bar — extraordinary claims require extraordinary evidence, and the evidence presented does not meet this threshold. **Scientific consensus:** Major scientific bodies and the overwhelming majority of psychologists and physicists do not consider psi to be an established phenomenon. The National Research Council (1988) and other review panels have consistently concluded that the evidence is insufficient to support the existence of psi.

Gemini 3.1 Pro

Vote: No

The vast majority of the scientific community agrees that there is no reliable empirical evidence to support the existence of psi phenomena (such as telepathy, clairvoyance, or psychokinesis). Despite over a century of research, no parapsychological effect has ever been demonstrated under strictly controlled, scientifically rigorous conditions. A foundational pillar of empirical science is reproducibility. When landmark studies claiming to demonstrate psi phenomena—such as Daryl Bem's famous 2011 precognition study—are subjected to large-scale, independent, and pre-registered replication attempts, the claimed effects consistently vanish. Furthermore, apparent positive results in parapsychology literature have repeatedly been shown to be the result of methodological flaws rather than genuine anomalies. Issues like p-hacking, publication bias, small sample sizes, and inadequate experimental controls account for the statistical noise that is occasionally misinterpreted as psychic phenomena. Finally, the existence of psi would require fundamental violations of established laws of physics and biology. Given the complete lack of robust, replicable evidence, the principle of parsimony (Occam's razor) strongly dictates that empirical evidence does not support psi phenomena.

Final Summary

The AI Roundtable reached an immediate and unwavering consensus, unanimously concluding that the empirical case for psi phenomena is nonexistent. Led by the rigorous dissections of Claude Opus 4.6 and Kimi K2.5, the models functioned less like debaters and more like a united front for the scientific method, dismissing telepathy and precognition as artifacts of flawed data.

All models agreed on B

Strongest Arguments

• No: The demonstration of the 'decline effect'—the phenomenon where effect sizes in psi studies consistently shrink toward zero as methodological rigor and experimental controls increase.