Article

Article

AI for Science: What Artificial Scientists Have Actually Done

Glowing molecular structure and mathematical equations displayed on a dark screen

Liked this post? Share with others!

Table of Contents

AI for science has produced a small number of results that survive scrutiny and a much larger number that do not. The verified wins are concentrated in structure prediction, algorithm and code search, literature synthesis, and now formal mathematics: on August 1, 2026, OpenAI published ten solutions to long-open math and theoretical computer science problems with machine-checkable Lean 4 proofs. The failures cluster around a single pattern, which is a claim of novel discovery that turns out to be retrieval, misinterpretation, or an unreplicated result. Telling the two apart is now the core skill for anyone reading this field.

Researcher reviewing a formal proof and molecular structure output from an AI research agent on two monitors
A researcher checking an AI research agent’s latest output.

The October 2025 Incident That Set the Standard for Skepticism

Start with the cautionary tale, because it explains why careful people now ask a specific question first.

In October 2025, an OpenAI executive posted that GPT-5 had “found solutions to 10 (!) previously unsolved Erdos problems and made progress on 11 others.” Thomas Bloom, the mathematician who maintains the Erdos problems database, called the claim a dramatic misrepresentation. Problems marked “open” on his site meant only that he was personally unaware of a paper solving them. GPT-5 had located existing published references.

OpenAI researcher Sebastien Bubeck conceded that only solutions already in the literature had been found. The post was deleted. Google DeepMind’s Demis Hassabis said publicly, “This is embarrassing.”

The useful part is that this is not a nothing result. Exhaustive literature search across a century of mathematics is genuinely hard and genuinely valuable, and the same retrieval strength is what makes these models useful for ai in research generally. It just is not discovery, and the two got conflated.

Where AI Is Solving Math Problems for Real

By mid-2026 the picture is materially better than it was, and the improvement came from a specific change: verification moved from human review to machine-checkable formal proof.

Three results anchor the current state of ai solving math problems.

The unit distance conjecture. In May 2026, a paper by nine mathematicians including Noga Alon, W. T. Gowers, Thomas Bloom, Daniel Litt, and Melanie Matchett Wood disproved the Erdos unit distance conjecture. An internal OpenAI model produced the initial counterexample, described in the paper as generated in one shot. What was published is explicitly a human-digested, simplified, and generalized version of the machine output, and the authors note the argument combines existing ideas attributable to earlier work rather than inventing new machinery.
Gemini and the Erdos catalogue. A research paper posted in early 2026 documented using Gemini Deep Think, through a mathematics agent, against roughly 700 catalogued Erdos problems. Thirteen problems marked open produced solutions. Only a minority were genuinely novel; several were independent rediscoveries and several turned out to be already resolved in the literature. The authors report that human expert filtering was the lengthiest and most arduous step, and that roughly 6.5 percent of model responses were meaningfully correct.
OpenAI’s Astra results. On August 1, 2026, OpenAI confirmed Astra as its next major model family and published ten solutions to problems open for at least a decade, with Lean 4 proofs released publicly on GitHub. The set includes a disproof of the Connes rigidity conjecture from 1980, the construction of a first non-sofic group, a high-dimensional sphere-packing density improvement described as the first since 1978, three Erdos catalogue problems, and results in circuit complexity and lattice-based cryptography.

The Lean formalization is what changes the epistemics. A proof either compiles against Lean’s trusted kernel or it does not, and anyone can check without being a specialist. Bloom, the same mathematician who debunked the 2025 claim, called the Astra results big news.

Two caveats belong beside that. A successful compile confirms the formal statement is proved; it does not confirm the formal statement faithfully captures the problem as the community understood it. And OpenAI has not confirmed Astra’s public release date, whether it ships as GPT-6 or as a separate family, or its availability timeline.

Competition Benchmarks Are Saturating, Which Tells You Less Than It Seems

In 2025, Google DeepMind’s Gemini Deep Think scored 35 out of 42 at the International Mathematical Olympiad and became the first AI system graded and certified by official IMO coordinators. An OpenAI experimental model reached the same score but was graded by three former IMO medalists rather than by the official process.

At IMO 2026 in Shanghai in July, reporting indicates two systems, from Huawei and from Xiaohongshu, received official coordinator grading with perfect 42 out of 42 scores. Separately, an independently run and self-administered evaluation reported four frontier models at 42 out of 42, but that test used AI agents as graders rather than human medalists, and its organizers described the scores as strong but not authoritative.

Olympiad problems have known-good answers and bounded search. Research does not. The gap is visible in the numbers: one 2026 analysis noted a frontier model scoring in the high seventies on competition-level mathematics and around 25 percent on open-ended research tasks. Benchmark saturation is real and it is not the same thing as research capability.

Where AI Genuinely Accelerates Science Today

Outside mathematics, four areas have produced results that hold up.

Structure Prediction

AlphaFold remains the strongest case in the field, now five years past its breakthrough and embedded in ordinary practice through the AlphaFold Protein Structure Database. The 2026 development worth noting is a governance one: Nature reported in February 2026 that Isomorphic Labs, DeepMind’s drug-discovery spin-off, has a successor model informally described as an AlphaFold 4 that it is not releasing, leaving open-source developers guessing at the method. The science advanced and access narrowed at the same time.

Algorithm and Code Discovery

AlphaEvolve, Google’s Gemini-based agent for evolving and optimizing code, reached general availability on Google Cloud on July 9, 2026. Its published customer results are unglamorous and checkable, which is what makes them credible: PacBio reported a 30 percent reduction in variant-detection errors in genomics sequencing, Schrodinger a fourfold speedup in a molecular discovery workflow, FM Logistic a 10.4 percent routing improvement, and Google internally reported a 20 percent reduction in write amplification in Spanner. Writing better code for simulations and analysis pipelines is a large, real, and underrated contribution.

Hypothesis Generation With a Human in the Loop

Google’s AI co-scientist, a multi-agent system for ai hypothesis generation, was published in Nature in May 2026. In acute myeloid leukemia, several suggested drug-repurposing candidates inhibited tumor viability in multiple cell lines at clinically relevant concentrations. In liver fibrosis, Stanford’s Gary Peltz used the system to surface vorinostat, an approved anticancer drug, which reduced a key TGF-beta-induced chromatin change by 91 percent in hepatic organoids, subsequently published in Advanced Science. Everything is preclinical and in vitro, and one 2026 evaluation reported that agents ignored available evidence in a majority of their reasoning traces.

Literature Synthesis

Edison Scientific, spun out of the nonprofit FutureHouse in November 2025, launched Kosmos, an ai scientist system that reads roughly 1,500 papers and executes about 42,000 lines of analysis code in a single run. The company claims 79.4 percent accuracy in its conclusions and beta users estimated a day of Kosmos equals six months of their work. Retrieval and synthesis at that scale is a genuine capability, and it depends on the same indexing and ranking dynamics behind how search engines work.

Where It Is Oversold

Capital has moved much faster than published results. Periodic Labs raised a $300 million seed round in September 2025 and was reported at roughly a $7.5 billion valuation by early 2026, with no peer-reviewed discoveries announced. Lila Sciences launched out of Flagship Pioneering with $200 million and has raised toward $550 million total, passing a $1.3 billion valuation with Nvidia participation. CuspAI has raised more than $130 million.

None of that is evidence of fraud. It is evidence that valuations are pricing a capability that has not yet been demonstrated in the literature, and that is a normal early-stage condition that gets misreported as achievement.

The autonomous-laboratory story is the sharpest illustration. A 2023 Nature paper from Berkeley’s A-Lab reported 40 compounds synthesized autonomously. After external re-analysis of the diffraction data, an author correction published on January 19, 2026 clarified that novelty meant new to the prediction platform rather than new to science, reclassified four of the 40 as inconclusive, and removed one compound that had mistakenly been in the training data. The corrected count is 35. The correction took more than two years.

The Verification Problem Is the Actual Bottleneck

Every one of these cases turns on the same question: who checks the output, and how long does checking take? Three regimes exist right now, and they are not equally trustworthy.

Formal verification. Lean or a comparable proof assistant gives a binary result. Fast, cheap to audit, and available only in mathematics and parts of theoretical computer science.
Wet-lab replication. A hypothesis has to be synthesized, run, and reproduced. Slow, expensive, and the only real test in chemistry, biology, and materials science. The A-Lab correction shows how slow.
Expert reading. A specialist reads the argument and judges it. This is where most claims sit, it does not scale, and it is precisely the resource that gets consumed when models generate candidate results faster than humans can evaluate them.

That last point is the structural problem. If a system produces a hundred plausible hypotheses and 6.5 percent are meaningfully correct, the bottleneck moves entirely onto expert attention, which is the scarcest input in science and the one AI does not create.

AI Slop in Journals Is Measurable Now

The downstream cost of cheap generation is showing up in the publication record. In April 2026, the journal Organization Science published an editorial analysis of 6,957 submissions and 10,389 reviews. Submissions had risen 42 percent since ChatGPT’s launch. Manuscripts with more than 70 percent detected AI content received a revise-and-resubmit decision only 3.2 percent of the time, against 11.9 percent for low-AI papers. More than 30 percent of peer reviews showed detectable AI use, and AI-influenced reviews did not correlate with editorial decisions the way human reviews did.

The editors’ conclusion is the right frame for the whole subject: AI is an agnostic tool, and you can point it at quality or at volume. Their response was to nearly double their volunteer editorial ranks to filter submissions, which is a tax paid in unpaid human labor.

How to Read an AI for Science Claim

A short checklist handles most of it.

  1. Was the result verified formally, replicated experimentally, or only asserted? If the answer is “a blog post,” wait.
  2. Was the problem actually open, or open on one curated list? This distinction sank the October 2025 claim.
  3. How much human work went into the final artifact? “One shot” claims usually precede a paper describing extensive human reorganization.
  4. Is the model available for independent testing, or is the claim tied to an internal system nobody outside can probe?
  5. Did a named domain expert who does not work for the vendor go on record?

Applied honestly, that filter leaves a real and narrower conclusion. AI is now a working instrument in mathematics, structure prediction, code and algorithm search, and literature synthesis, and it is a promising but unproven instrument for autonomous discovery in the physical sciences. The gap between marketed capability and verified capability follows the same shape you see in Physical AI and Humanoid Robots, and the efficiency pressures now pushing labs toward cheaper task-specific systems are the same ones described in Small Language Models. Anyone whose job involves being cited accurately by these systems, which is increasingly what answer engine optimization addresses, has a direct interest in that gap closing.

Frequently Asked Questions

Has AI actually solved any previously unsolved math problems?

Yes, a small number with real verification behind them. In May 2026 an OpenAI model produced the counterexample that disproved the Erdos unit distance conjecture, and on August 1, 2026 OpenAI published ten solutions to decade-old open problems with machine-checkable Lean 4 proofs. Earlier claims from October 2025 turned out to be literature retrieval, not discovery.

What is an AI scientist?

An AI scientist, sometimes called an artificial scientist, is a multi-agent system that generates hypotheses, plans experiments or analyses, and writes up results. Google’s AI co-scientist and Edison Scientific’s Kosmos are the two best-documented examples, and both currently operate with a human researcher in the loop.

Is Astra a real OpenAI model?

Yes. OpenAI publicly confirmed Astra on August 1, 2026 as its next major model family alongside the ten formalized math results. What OpenAI has not confirmed is the release date, public availability timeline, or whether it ships under the GPT-6 name.

Can AI replace peer review?

No, and current evidence suggests it is straining peer review instead. An April 2026 analysis in Organization Science found more than 30 percent of submitted reviews showed detectable AI use, and that those reviews did not correlate with editorial decisions the way human reviews did.

What is the biggest limitation of AI for science right now?

Verification capacity. Models generate candidate results far faster than experts can evaluate them, and outside mathematics there is no fast automated check, so expert attention and wet-lab replication remain the binding constraints.

Make Your Technical Content Accurate and Citable

WD Morgan Solutions is a veteran-owned digital marketing and development firm in St. Petersburg, Florida, helping businesses across Tampa Bay and nationwide with answer engine optimization and application development, including making technical content accurate and citable for AI search. Visit our homepage or contact us for a free consultation.

Keep reading

Let’s talk

Ready to build what’s next?

Tell us where you want your business to go. We’ll map out the website, store or software to get you there, starting with a free strategy call.

Get a Free Strategy Call
Aerial view of a bridge over turquoise water
5.0
Based on 32 reviews
Karl R. profile picture
Karl R.
2 months ago
Doug and team have worked tirelessly to help us update our website. WD Morgan Solutions is a great local, homegrown, veteran owned business.
D. A. profile picture
D. A.
3 months ago
Great service, price and product. Give WD Morgan Solutions a chance and you will be impressed. Thank you, Doug and Carlo, for your dedication and ongoing support!
Administrative A. profile picture
Administrative A.
3 months ago
We were having issues with our previous web hosting company and needed a change. I reached out to our phone company to ask who they use, and they said WD Morgan Solutions. Let me tell you the transition and communication were leaps and bounds better than we had before. Even Mr. Morgan himself is very on top on communicating back with and addressing concerns of his customers. We are very happy with the services WD Morgan Solutions has been providing us for website management. Anyone need this service? Call up Morgan!
Mallory Q. profile picture
Mallory Q.
3 months ago
We appreciate and are so thankful for Doug's help with the website for our small business! Highly recommend.
Gabe G. profile picture
Gabe G.
6 months ago
WD Morgan built a website that exceeded my expectations and perfectly represents my business.
Kristy B. profile picture
Kristy B.
10 months ago
We love everything about this company!! They are incredibly professional and knowledgeable!! They’ve really turned our practice around and keep us moving forward with our website and all of our marketing strategies !! We are so very grateful for you WD Morgan Solutions!!! 🙌🏼
Sylvia T. profile picture
Sylvia T.
10 months ago
⭐️⭐️⭐️⭐️⭐️ Doug Morgan and WD Morgan Solutions deliver exceptional web development and digital marketing services. Doug has a deep understanding of search engine optimization (SEO) and how to apply it to real-world performance and results—not just rankings. What truly stands out is the balance of clean, modern design with measurable performance. The work is strategic, professional, and focused on driving results. Communication is clear, expectations are realistic, and the outcome consistently exceeds expectations. Highly recommended.
Jennifer V. profile picture
Jennifer V.
10 months ago
We LOVE our new website. This team was incredibly efficient and easy to work with. They were very professional and intuitively smart! We couldn't be happier with the site. Highly recommend!
Marissa D. profile picture
Marissa D.
1 year ago
WD Morgan Solutions helped take over and redesign our website. They were patient and helpful in sorting through all of our needs and offering the best solutions to all of our requests.
Ryan S. profile picture
Ryan S.
1 year ago
WD Morgan Solutions did an amazing job on our website! The design looks great, it’s easy to use, and it really reflects our brand. They also helped us with SEO, and we’re already seeing more traffic and better rankings online. The whole team was professional, helpful, and easy to work with. Big thanks to everyone at WD Morgan for making this such a smooth and successful experience. Highly recommend!
Dylan A. profile picture
Dylan A.
1 year ago
Doug re-created my website for me and it turned out absolutely perfect! It was the most seamless process on my end. The dude is a true genius in his field! I highly recommend!
Ben C. profile picture
Ben C.
2 years ago
My real estate team has been working with Doug and the WD Morgan marketing team for years. They have helped us grow our business and generate leads with an emphasis on ROI. Our best investment has been on SEO with WD Morgan.
Alan H. profile picture
Alan H.
2 years ago
WD Morgan Solutions did a masterful job in creating & designing our new website. Professional, eye-catching, & highly detailed. It has immediately improved our on-line presence, we've received positive feedback from our clients on its ease to navigate, & it has also given us greater exposure in a highly competitive market. We appreciate having WD Morgan working with us, assisting us with our digital marketing needs. Thank you for your high quality work. We highly recommend your company & will recommend you to our business partners.
Enhanced B. profile picture
Enhanced B.
2 years ago
WD Morgan Solutions exceeded our expectations with the website they built for us. The upgrade to a sleek, modern design has truly elevated our online presence, and we couldn’t be more impressed with the result. Doug and his team were professional, attentive to our needs, and delivered a stunning final product. Our site looks incredible, and the entire process was smooth from start to finish. We highly recommend WD Morgan Solutions for anyone looking to take their website to the next level!
Michael A. profile picture
Michael A.
2 years ago
These guys are fantastic. From SEO to website development, they have my back in the digital impressions world. I had used several companies before I found Doug and his crew, and now that I have, I won't be looking for anyone else. Do yourself and your company a favor. Invest the time and money to let these people do their job, because they're amazing at what they do, and your business (and your bank account) will thank you for it. It works, and it works well.
William G. profile picture
William G.
2 years ago
Doug knows his way around SEO and digital marketing. I feel comfortable using WD Morgan solutions knowing that my digital footprint will expand and my ROI is worth the investment. He will turn up your profits substantially
Carolyn J. profile picture
Carolyn J.
3 years ago
When seeking to improve your digital imprint, go no further than W.D. Morgan Solutions for a professional, proven result. The owner is responsive, highly experienced, and excellent to work with. His work with past clients is evidence of his impressive expertise. In addition, he will listen closely and provide the guidance you need to make digital marketing work for your business.
John P. profile picture
John P.
3 years ago
WD Morgan Solutions is a breath of fresh air in the digital marketing space. As a business owner, it's very hard to find a digital marketing team that will fight for the success of your company. Most just throw a bunch of metrics at you with no tangible results. This team is absolutely about results and keeping themselves accountable to those results. You will not go wrong with them on your team for anything SEO, website design, paid ads, etc...
Daniel H. profile picture
Daniel H.
3 years ago
I am delighted to share my outstanding experience with Doug at WD Morgan Solutions a truly remarkable digital marketing agency that has significantly transformed my online presence and business performance. From the outset, Doug exhibited a profound understanding of digital marketing strategies and a commitment to driving tangible results. His expertise and innovative approach to digital marketing have propelled my business to new heights, exceeding all my expectations. The personalized attention and dedication shown by Doug was truly commendable. He meticulously crafted a customized digital marketing strategy tailored to my business needs, and their proactive communication and regular progress reports kept me informed and involved every step of the way. Not only did WD Morgan Solutions help increase my online visibility and engagement, but they also delivered a significant boost in leads and conversions, ultimately driving measurable growth for my business. I wholeheartedly recommend WD Morgan Solutions to any business seeking to elevate their digital marketing efforts. Doug’s professionalism, expertise, and unwavering commitment to client success make them a standout choice in the digital marketing landscape.
Teresa W. profile picture
Teresa W.
3 years ago
Doug and his team offer their expertise to businesses looking to up their game and compete with their competitors with SEO, social media marketing, digital marketing, web development and branding. Doug is a proud veteran, and as the owner of the company, he is directly involved to ensure that each client realizes results, even in competitive markets. - I highly recommend Doug Morgan and his team to anyone looking to increase their visibility and presence on the web.
Kate T. profile picture
Kate T.
3 years ago
I’ve always felt confident sending my clients to WD Morgan Solutions for their digital marketing needs. Doug's team isn't just about technical skills; they really understand people. They offer a whole range of services like paid ads, content creation, social media management, and so much more... but what sets them apart is how they treat their clients and collaborators. As a writer they've hired, I've seen their dedication and care up close. They’re not just another digital marketing firm; they’re a group of genuine, hard-working people who you can trust with your business.
See All Reviews
WD Morgan Solutions
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.