6,750 claims · page 18 of 68
| Verdict | Person | Kind | Said | Claim | Topic |
|---|---|---|---|---|---|
| Mostly false | Marc Andreessen | Factual | 26 Jan 2025 | Getting Amish voters to the polls was one of the reasons Trump won the 2024 election. Marc Andreessen · Factual · 26 Jan 2025 | AI policy and governance |
| Mostly true | Marc Andreessen | Factual | 26 Jan 2025 | Incumbent universities, through accreditation, are in complete control of who gets access to federal student loan money. Marc Andreessen · Factual · 26 Jan 2025 | Other |
| Mostly true | Marc Andreessen | Factual | 26 Jan 2025 | Marc Andreessen does not speak for DOGE and is not in charge of DOGE. Marc Andreessen · Factual · 26 Jan 2025 | AI policy and governance |
| Mostly true | Marc Andreessen | Factual | 26 Jan 2025 | The United States is the main place where the world's best and brightest foreign students get educated. Marc Andreessen · Factual · 26 Jan 2025 | AI, the economy and work |
| Unresolvable | Marc Andreessen | Factual | 26 Jan 2025 | Most of the Republican Party's institutional base were Bush Republicans. Marc Andreessen · Factual · 26 Jan 2025 | AI policy and governance |
| Unresolvable | Marc Andreessen | Factual | 26 Jan 2025 | Around 2012-2013, internet companies rapidly pivoted toward social media censorship and content banning. Marc Andreessen · Factual · 26 Jan 2025 | Philosophy, ethics and society |
| Unresolvable | Marc Andreessen | Prediction | 26 Jan 2025 | Andreessen said he thinks it may be that, with the continued growth of Substack and other new platforms, a tremendous amount of previously forbidden conversations can now shift back into public view, though he did not know that everything would. Marc Andreessen · Prediction · 26 Jan 2025 | Philosophy, ethics and society |
| Unresolvable | Marc Andreessen | Prediction | 26 Jan 2025 | Andreessen said he thinks that if taxpayers come to understand the level of crisis, the funding will evaporate. Marc Andreessen · Prediction · 26 Jan 2025 | Other |
| Mostly true | Marc Andreessen | Factual | 26 Jan 2025 | The US is adding a trillion dollars to the national debt every 100 days, and it is now passing the size of the Defense Department budget and compounding. Marc Andreessen · Factual · 26 Jan 2025 | Economics and markets |
| Mixed | Marc Andreessen | Factual | 26 Jan 2025 | The system that has been built for these people is paid for virtually 100% by taxpayer money. Marc Andreessen · Factual · 26 Jan 2025 | Economics and markets |
| Unresolvable | Marc Andreessen | Factual | 26 Jan 2025 | Marc Andreessen said he politely declined to design the 'breast detection algorithm' he had imagined he would have to build. Marc Andreessen · Factual · 26 Jan 2025 | Their own work or organisation |
| Mostly true | Marc Andreessen | Factual | 26 Jan 2025 | Marc Andreessen has made the argument in favour of high-skilled immigration and H-1B visas many times. Marc Andreessen · Factual · 26 Jan 2025 | AI, the economy and work |
| Mostly true | Marc Andreessen | Factual | 26 Jan 2025 | Marc Andreessen joined the Facebook board in 2007. Marc Andreessen · Factual · 26 Jan 2025 | Their own work or organisation |
| Unresolvable | Marc Andreessen | Prediction | 26 Jan 2025 | If the debt problem is not fixed, the US will at some point enter a hyperinflationary spiral and become like Argentina or Brazil. Marc Andreessen · Prediction · 26 Jan 2025 | Economics and markets |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | For the next period, running advanced reasoning models will take a lot of compute. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI compute, hardware and data |
| Own records only | Mark Zuckerberg | Factual | 10 Jan 2025 | Zuckerberg said that it was probably over about a few years that he started coming to the conclusion that Meta would need to change something about a program he described as destroying trust, especially in the United States. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Brain-interface technology will first be used by people with severe conditions before being used for gaming. Mark Zuckerberg · Prediction · 10 Jan 2025 | Technology |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | AI engineers will augment the people working on code. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI, the economy and work |
| Resolved true | Mark Zuckerberg | Factual | 10 Jan 2025 | Zuckerberg controls the majority of the voting power at Meta. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Zuckerberg guessed that if AI is open source and widely available, people will all kind of have superpowers, rather than a runaway AI emerging. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI policy and governance |
| Own records only | Mark Zuckerberg | Factual | 10 Jan 2025 | Meta added disappearing messages to its messaging products partly for security reasons. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Mixed | Mark Zuckerberg | Prediction | 10 Jan 2025 | Requiring more confidence for takedowns will dramatically reduce the number of people whose Meta accounts are removed by mistake. Mark Zuckerberg · Prediction · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | At every step of human technological progress, people thought the technology being developed would make people obsolete. Mark Zuckerberg · Factual · 10 Jan 2025 | AI, the economy and work |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | Meta put in place a fact-checking program that empowered third-party fact checkers to mark content false, in response to critiques about misinformation. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | If AI is locked down, only big AI companies and the Chinese government (which steals it from them) would have access to it. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI policy and governance |
| Mostly true | Mark Zuckerberg | Prediction | 10 Jan 2025 | Changing Meta's content filters to require higher confidence and precision will reduce the vast majority of Meta's censorship mistakes. Mark Zuckerberg · Prediction · 10 Jan 2025 | Their own work or organisation |
| Mostly true | Mark Zuckerberg | Factual | 10 Jan 2025 | Someone in Pakistan tried to get Zuckerberg sentenced to death, bringing a criminal proceeding over a Facebook picture of the prophet Muhammad. Mark Zuckerberg · Factual · 10 Jan 2025 | Philosophy, ethics and society |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Meta will go to court against government investigations and win all the cases because it follows the rules. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI policy and governance |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | After 2016 Zuckerberg gave too much deference to media who said Trump's election was due to misinformation. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | The AR demo Joe Rogan played used version 1 of Meta's neural interface. Mark Zuckerberg · Factual · 10 Jan 2025 | Technology |
| Own records only | Mark Zuckerberg | Factual | 10 Jan 2025 | Without Apple's rules and fees, Meta would make about twice as much profit, by Zuckerberg's rough calculation. Mark Zuckerberg · Factual · 10 Jan 2025 | Economics and markets |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | It will be a while before brain-interface technology is widely deployed. Mark Zuckerberg · Prediction · 10 Jan 2025 | Technology |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | Zuckerberg was worried from the beginning about Meta becoming the decider of what is true. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | Zuckerberg said he thinks that while they were pushing the vaccine program, they also tried to censor anyone who was arguing against it. Mark Zuckerberg · Factual · 10 Jan 2025 | AI policy and governance |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | Meta generally deferred to the government on some policies that Zuckerberg probably would not follow knowing what he knows now. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Over time, a lot of the code in Meta's apps, including its AI, will be built by AI engineers instead of people engineers. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI capabilities |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | A single VR player with a haptic suit will not be able to actually kick someone who is not physically there. Mark Zuckerberg · Prediction · 10 Jan 2025 | Technology |
| Resolved true | Mark Zuckerberg | Factual | 10 Jan 2025 | Before Meta updated its policies, a statement calling for women to be excluded from certain combat roles could not be posted on its platforms because it called for exclusion of a protected category. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | The number of hours per week people must work to get by will probably continue to shrink. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI, the economy and work |
| Mostly true | Mark Zuckerberg | Factual | 10 Jan 2025 | Apple built a protocol for connecting accessories to the iPhone that is not encrypted (plain text). Mark Zuckerberg · Factual · 10 Jan 2025 | Technology |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | In the future people will have AI coworkers. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI, the economy and work |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Jiu-jitsu will be the last thing that can be done in VR because it needs the other person's momentum. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI capabilities |
| Unresolvable | Mark Zuckerberg | Factual | 10 Jan 2025 | Zuckerberg was sympathetic to temporary speech curbs at the beginning of COVID. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Zuckerberg said he thought that probably in 2025 Meta and other companies working on AI would have an AI that can effectively act as a mid-level engineer writing code. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI capabilities |
| Mostly true | Mark Zuckerberg | Factual | 10 Jan 2025 | The body has redundant neural pathways to the hand because the ability to use the hand is very important. Mark Zuckerberg · Factual · 10 Jan 2025 | Technology |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | The U.S. government will probably need to help defend American tech companies against foreign governments. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI policy and governance |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Apple's second and third versions of the Vision Pro will probably be better than the first version. Mark Zuckerberg · Prediction · 10 Jan 2025 | Technology |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | People who are highly engaged in their work will accomplish far more than before thanks to the leverage of more technology. Mark Zuckerberg · Prediction · 10 Jan 2025 | AI, the economy and work |
| Resolved true | Mark Zuckerberg | Factual | 10 Jan 2025 | Zuckerberg said that there are emails and that the emails have been published. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Own records only | Mark Zuckerberg | Factual | 10 Jan 2025 | Zuckerberg communicates with only a few people over SMS, mostly using WhatsApp, Instagram Direct or Messenger. Mark Zuckerberg · Factual · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Zuckerberg said he thinks a whole new class of creators will become the new cultural elites that people look to. Mark Zuckerberg · Prediction · 10 Jan 2025 | Philosophy, ethics and society |
| Unresolvable | Mark Zuckerberg | Prediction | 10 Jan 2025 | Zuckerberg said he thinks Meta's current approach will be pretty durable going forward, because the company has been pressure-tested on these issues for about eight to ten years. Mark Zuckerberg · Prediction · 10 Jan 2025 | Their own work or organisation |
| Unresolvable | Jensen Huang | Factual | 7 Jan 2025 | Everything shown inside NVIDIA's digital twin is generated by AI. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Mixed | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA Cosmos is the world's first world foundation model. Jensen Huang · Factual · 7 Jan 2025 | AI capabilities |
| Resolved true | Jensen Huang | Factual | 7 Jan 2025 | On 2025-01-07, Jensen Huang announced NVIDIA's next-generation RTX Blackwell family, the GeForce RTX 50 series. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Unresolvable | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA's NV1 made it possible to have a game console in a PC. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Mostly false | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA DRIVE OS is the first software-defined programmable AI computer certified up to ASIL D, the highest automotive functional safety standard. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA fine-tuned Llama models and turned them into the Llama Nemotron suite of open models. Jensen Huang · Factual · 7 Jan 2025 | AI capabilities |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA Cosmos is trained on 20 million hours of video. Jensen Huang · Factual · 7 Jan 2025 | AI capabilities |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA's Blackwell chip is in full production as of January 2025. Jensen Huang · Factual · 7 Jan 2025 | AI compute, hardware and data |
| Mostly false | Jensen Huang | Promise | 7 Jan 2025 | NVIDIA's GB10-based Project DIGITS computer will be available around May 2025. Jensen Huang · Promise · 7 Jan 2025 | Technology |
| Own records only | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA concluded Llama models could be better fine-tuned for enterprise use. Jensen Huang · Factual · 7 Jan 2025 | AI capabilities |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA invented the programmable GPU in 1999, six years after the UDA work described just before in the speech, and this started more than 20 years of advances in the GPU. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Resolved true | Jensen Huang | Factual | 7 Jan 2025 | Toyota and NVIDIA are going to partner to create their next-generation autonomous vehicles. Jensen Huang · Factual · 7 Jan 2025 | Economics and markets |
| Unresolvable | Jensen Huang | Prediction | 7 Jan 2025 | The ChatGPT moment for general robotics is just around the corner. Jensen Huang · Prediction · 7 Jan 2025 | Technology |
| Unresolvable | Jensen Huang | Prediction | 7 Jan 2025 | Every robotics company will ultimately have to build three computers, one of them to train the AI. Jensen Huang · Prediction · 7 Jan 2025 | Technology |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | The RTX 50 series Blackwell GPU has 92 billion transistors, 4,000 TOPS and four petaflops of AI, three times the previous-generation Ada. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Resolved true | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA invented CUDA six years after 1999 (around 2005) to expose GPU programmability to a broad set of algorithms. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA invented neural texture compression and neural material shading. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | In the RTX 50 (Blackwell) generation the programmable shader can process neural networks. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Mostly false | Jensen Huang | Factual | 7 Jan 2025 | The RTX 5070 laptop priced at $1,299 has RTX 4090 performance at half the power. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Mostly true | Jensen Huang | Factual | 7 Jan 2025 | NVIDIA's AI Blueprints are completely open source. Jensen Huang · Factual · 7 Jan 2025 | Technology |
| Not yet due | Rodney Brooks | Prediction | 1 Jan 2025 | Robotaxi service, even with remote operators, will not be operating in 50 US cities by 2028. Rodney Brooks · Prediction · 1 Jan 2025 | Economics and markets |
| Mostly true | Rodney Brooks | Factual | 1 Jan 2025 | Zoox has purpose-built vehicles with no steering wheel or brake and accelerator pedals. Rodney Brooks · Factual · 1 Jan 2025 | Technology |
| Not yet due | Rodney Brooks | Prediction | 1 Jan 2025 | Humanoid robots will not take away jobs for decades. Rodney Brooks · Prediction · 1 Jan 2025 | AI, the economy and work |
| Not yet due | Rodney Brooks | Prediction | 1 Jan 2025 | US battery electric cars will not reach 30% of new car sales in 2027, given a Q3 2024 share of 8.9%. Rodney Brooks · Prediction · 1 Jan 2025 | Economics and markets |
| Unresolvable | Rodney Brooks | Prediction | 1 Jan 2025 | Boeing's Starliner may never fly again. Rodney Brooks · Prediction · 1 Jan 2025 | Economics and markets |
| Unresolvable | Rodney Brooks | Prediction | 1 Jan 2025 | Coding as a job is not about to be eliminated by AI. Rodney Brooks · Prediction · 1 Jan 2025 | AI, the economy and work |
| Unresolvable | Gary Marcus | Prediction | 1 Jan 2025 | OpenAI will continue in 2025 to preview products months or years before they are solid and widely available at an accessible price. Gary Marcus · Prediction · 1 Jan 2025 | AI capabilities |
| Resolved true | Gary Marcus | Prediction | 1 Jan 2025 | AI Safety Institutes will offer guidance in 2025 but have little legal authority to stop truly dangerous models. Gary Marcus · Prediction · 1 Jan 2025 | AI policy and governance |
| Unresolvable | Rodney Brooks | Prediction | 1 Jan 2025 | It would not be a huge surprise if Boeing's Starliner never flies again. Rodney Brooks · Prediction · 1 Jan 2025 | Other |
| Resolved true | Gary Marcus | Factual | 1 Jan 2025 | As of end 2024, AI model makers have very little moat, with many similar models and a price war. Gary Marcus · Factual · 1 Jan 2025 | Economics and markets |
| Unresolvable | Gary Marcus | Prediction | 1 Jan 2025 | No 'GPT-5 level' model, a huge across-the-board leap by community consensus, may appear throughout 2025. Gary Marcus · Prediction · 1 Jan 2025 | AI timelines and forecasting |
| Not yet due | Rodney Brooks | Prediction | 1 Jan 2025 | Brooks says he "got suckered by his hype" and that the development he discusses, which the surrounding text links to a "taxi" service in 50 of the 100 biggest US cities, is unlikely to happen in the first half of this century. Rodney Brooks · Prediction · 1 Jan 2025 | Economics and markets |
| Not yet due | Rodney Brooks | Prediction | 1 Jan 2025 | Brooks said he is quite confident that SpaceX will not fly 400 Starship launches in the four years 2025 through 2028. Rodney Brooks · Prediction · 1 Jan 2025 | Economics and markets |
| Unresolvable | Rodney Brooks | Prediction | 1 Jan 2025 | The imagined capabilities of many of the technologies in question will not arrive in the time frames most of the technology and prognosticator class expect. Rodney Brooks · Prediction · 1 Jan 2025 | AI timelines and forecasting |
| Unresolvable | Gary Marcus | Factual | 1 Jan 2025 | Marcus coined in 2022 the notion that AI might be 'hitting a wall', which later became widespread in outlets such as CNN, The Verge and the Wall Street Journal. Gary Marcus · Factual · 1 Jan 2025 | Their own work or organisation |
| Mostly true | Gary Marcus | Factual | 1 Jan 2025 | OpenAI has returned no profits, does not expect profits until at least 2029, has a burn rate in the billions, and lacks a clear technical moat. Gary Marcus · Factual · 1 Jan 2025 | Economics and markets |
| Resolved true | Gary Marcus | Prediction | 1 Jan 2025 | Profits from AI models will remain modest or nonexistent in 2025. Gary Marcus · Prediction · 1 Jan 2025 | Economics and markets |
| Resolved true | Gary Marcus | Prediction | 1 Jan 2025 | Lack of reliability will continue to afflict generative AI in 2025. Gary Marcus · Prediction · 1 Jan 2025 | AI capabilities |
| Resolved true | Gary Marcus | Prediction | 1 Jan 2025 | Artificial general intelligence will not be achieved in 2025. Gary Marcus · Prediction · 1 Jan 2025 | AI timelines and forecasting |
| Resolved true | Rodney Brooks | Factual | 1 Jan 2025 | Boeing's Starliner did not bring its crew back from space as expected. Rodney Brooks · Factual · 1 Jan 2025 | Economics and markets |
| Unresolvable | Rodney Brooks | Prediction | 1 Jan 2025 | Scientists are not about to be replaced by AI programs. Rodney Brooks · Prediction · 1 Jan 2025 | AI, the economy and work |
| Not yet due | Rodney Brooks | Prediction | 1 Jan 2025 | The aircraft type whose last unit was built in 1963 is expected to keep flying until at least 2040, and perhaps longer. Rodney Brooks · Prediction · 1 Jan 2025 | Technology |
| Resolved true | Gary Marcus | Prediction | 1 Jan 2025 | Hallucinations will continue to afflict generative AI in 2025. Gary Marcus · Prediction · 1 Jan 2025 | AI capabilities |
| Resolved true | Gary Marcus | Prediction | 1 Jan 2025 | Less than 10% of the workforce will be replaced by AI in 2025, probably less than 5%. Gary Marcus · Prediction · 1 Jan 2025 | AI, the economy and work |
| Unresolvable | Rodney Brooks | Prediction | 1 Jan 2025 | Humans are not on the verge of being replaced and eliminated from white-collar or blue-collar jobs by AI and robots. Rodney Brooks · Prediction · 1 Jan 2025 | AI, the economy and work |
| Unresolvable | Rodney Brooks | Factual | 1 Jan 2025 | 40 companies hold permits to test autonomous driving in California, and after Cruise's demise Waymo is the only company trying to make an actual go of a digital taxi service in the US. Rodney Brooks · Factual · 1 Jan 2025 | Economics and markets |
| Resolved true | Rodney Brooks | Factual | 1 Jan 2025 | Despite further thruster anomalies, Starliner would have landed safely had the crew been aboard when it returned empty in 2024. Rodney Brooks · Factual · 1 Jan 2025 | Economics and markets |
| Unresolvable | Gary Marcus | Prediction | 1 Jan 2025 | Copyright lawsuits over generative AI will continue throughout 2025. Gary Marcus · Prediction · 1 Jan 2025 | AI policy and governance |
“The lack of reliability will continue to haunt generative AI.”
Gary Marcus, Substack author 1 January 2025 25 AI Predictions for 2025, from Marcus on AI (Marcus on AI)
Lack of reliability will continue to afflict generative AI in 2025.
Resolved true The record shows it happened or was so.
What would count as true or false
The claim is that unreliability would remain a significant, widely recognised problem for generative AI during 2025. Unreliability here means hallucinations, factual errors, and inconsistent or unpredictable outputs. True: through 2025, credible sources document ongoing reliability problems in leading generative AI systems. Examples include hallucination benchmarks showing non-trivial error rates, prominent publicised failures such as fabricated legal citations or erroneous outputs, and industry or academic reports naming reliability as a key barrier. Mostly true: such problems persist but are widely reported as substantially reduced. Mixed: the evidence is split, with major systems widely treated as reliable for most uses but notable failures still occurring. Mostly false: credible evidence shows reliability largely stopped being a significant concern for leading systems, with only isolated incidents. False: by 2025 leading generative AI is broadly documented as reliable, and reliability is no longer a recognised problem.
Evidence
- AI chatbots unable to accurately summarise news, BBC finds BBC News · 11 February 2025
“It found 51% of all AI answers to questions about the news were judged to have significant issues of some form.”
- AI chatbots fail at accurate news, major study reveals Deutsche Welle (DW) · 21 October 2025
“the study found that almost half of all answers had at least one significant issue, while 31% contained serious sourcing problems and 20% contained major factual errors.”
- How Strong Is AI When Hallucinations Haunt? G2 · 2 April 2025
“On average, reviews mentioning hallucinations dropped from 35% in March 2025 to 8.3% in October 2025, according to G2 review data.”
- AI Hallucinations May Soon Be History Inside Higher Ed · 1 October 2025
“While more advanced models may reduce the frequency of obvious factual mistakes, the issue persists in more subtle forms.”
Reasoning
Independent studies from 2025 document ongoing reliability problems in leading chatbots. The BBC found that 51% of AI answers about the news had significant issues (E1). A major study reported by DW in October 2025 found that almost half of answers had at least one significant issue, and 20% contained major factual errors (E2). G2 review data shows mentions of hallucinations falling over 2025 (E3), and Inside Higher Ed says the problem persists in subtler forms (E4). Taken together, unreliability remained a widely recognised problem through 2025, which meets the criteria for true.
Cite this claim
Gary Marcus, “Lack of reliability will continue to afflict generative AI in 2025.” (said 1 January 2025). Verdict: resolved true. Reliability API, claim c_4eeff44d9f40, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_4eeff44d9f40
Verdicts are made automatically; how verdicts and scores are made. The same data is on Hugging Face under CC BY 4.0.