iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
The three-year forecast in this headline’s source came from Sam Altman, OpenAI’s co-founder and CEO—not a TikTok CEO. In testimony to the U.S. Senate Committee on Commerce, Science, and Transportation on May 8, 2025, Altman said AI might help unlock scientific breakthroughs by designing experiments, with climate, health, and national security among the areas he named. That was a prediction, not evidence that such breakthroughs had occurred.
Who made the forecast?
Altman told the Senate committee, “I believe we’ll see many major advances over the next three years, but here are some examples.” The testimony then gave a timeline: sophisticated software-engineering tools and agents in 2025, possible AI-designed experiments leading to breakthroughs in 2026, and AI-powered robotics affecting physical productivity in 2027. These were expectations expressed in May 2025, not confirmed outcomes. The matching testimony is available in a transcript hosted on Scribd.
The source does not identify Altman as a TikTok executive. The headline’s attribution appears to be mistaken; the relevant speaker was OpenAI’s CEO.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →What did Altman predict about science?
His specific example was that AI could help scientists design experiments to address difficult problems. He said: “In 2026, AI may unlock a new wave of scientific breakthroughs by designing experiments to tackle America’s toughest challenges in climate, health, and national security.”
#1 Best Overall
The wording is deliberately prospective: AI may help, and the proposed mechanism is experiment design. The testimony does not say that AI had already produced the breakthroughs, define what would count as “major,” or set a measurable threshold for judging whether the forecast succeeded.
How to interpret the timeline now
Altman named 2026 as a possible year for AI-designed experiments to contribute to breakthroughs. Because the forecast was made in May 2025, that target year has arrived and is nearing its end as of October 2026. The cited material does not establish whether the prediction was fulfilled, so it should not be reported as a verified result. A sound assessment would require identifying specific scientific advances, showing AI’s contribution to the experiments, and deciding in advance what qualifies as a major breakthrough.
Rank #2
The testimony also mentioned software agents in 2025 and robotics in 2027. Those are separate parts of Altman’s anticipated progression, not evidence for the science forecast.
What later comments add—and what they do not
Altman’s remarks at a June 2025 event
At a separate event recorded by Reuters Connect on June 2, 2025, Altman said he would bet that limited cases of agents helping discover new knowledge might appear the following year. He also described a future “AI agent that can go discover new science” as a significant moment. These comments provide context for his thinking about AI agents; they are not independent confirmation of the Senate prediction. Reuters Connect event record.
Demis Hassabis’s separate view
In a May 2026 ABC News interview, Google DeepMind CEO Demis Hassabis described AI as a tool for scientific and medical discovery, mentioning disease, energy, and environmental challenges. His outlook is a separate perspective, not corroboration that Altman’s forecast came true. ABC News interview.
Investment figures are not a measure of scientific results
A 2025 Axios roundup discussed AI-related activity in health, genomics, and robotics. It reported, citing the Center for Strategic and International Studies, $3.3 billion in U.S. federal non-defense AI research and development investment for fiscal year 2025, and more than $109 billion in private-sector investment in 2024. Those figures describe investment, not breakthroughs or the accuracy of Altman’s prediction. The federal estimate’s definitions and methodology should be checked against the CSIS source before being used as a standalone statistic. Axios roundup.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What counts as evidence of an AI-driven breakthrough?
For any claimed result, distinguish a model that suggests an idea from a result validated by scientific work. Relevant questions include:
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches- Did AI materially help formulate or design the experiment, or was it used only for routine analysis?
- Were the methods and findings independently reviewed or replicated?
- Is the result a demonstrated advance in science or medicine, rather than an investment announcement or a company’s description of its work?
- Does the evidence connect the result to the timeframe and criteria in the original forecast?
In September 2026, OpenAI published a company account of internal AI-assisted research and a reported solution to the Navier–Stokes Millennium Prize Problem. The account says researchers still set priorities and assess results. It is a company report and does not independently validate Altman’s timeline; the mathematical claim should be assessed through the linked research material before being treated as established. OpenAI’s account.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
