AI is more efficient, more accessible and more affordable; in addition, Asians are more optimistic about AI.AI is more efficient, more accessible and more affordable; in addition, Asians are more optimistic about AI.

Highlights of Stanford HAI's 2025 Artificial Intelligence Index Report

2025/04/14 16:25

Author: Stanford HAI (Stanford Artificial Intelligence Institute)

Compiled by: Felix, PANews

Stanford HAI recently released the 456-page "Artificial Intelligence Index Report 2025". Here are some key points about artificial intelligence trends:

1. AI is becoming much more powerful than imagined

In the new benchmarks MMMU, GPQA, and SWE-bench, AI performance improved significantly: scores increased by 18.8%, 48.9%, and 67.3%, respectively. In addition to the benchmarks, AI systems made significant progress in generating high-quality videos, and in some cases, large language models (LLMs) even surpassed humans in timed programming tasks.

Note:

MMMU is a novel, carefully designed benchmark for multi-disciplinary multimodal understanding and reasoning at the university level, aiming to evaluate the expert-level multimodal understanding capabilities of underlying models on a wide range of tasks.

GPQA is a challenging dataset consisting of 448 high-quality and difficult multiple-choice questions written by experts in different fields. Experts who hold or are pursuing a PhD in the corresponding field achieve only 65% accuracy, while highly skilled non-expert verifiers achieve only 34% accuracy despite spending an average of more than 30 minutes and having unlimited access to the Internet.

SWE-bench is a benchmark for evaluating the performance of Large Language Models (LLMs) on real-world software questions collected from GitHub.

Highlights of Stanford HAI's 2025 Artificial Intelligence Index Report

2. AI is more efficient, accessible, and affordable

Smaller AI models with fewer parameters are becoming increasingly powerful: in just two years, the number of parameters has been reduced by about 100 times, while still scoring over 60% on the Massive Multi-Task Language Understanding (MMLU) test.

The gap between open source and closed source models is also narrowing, with the performance gap falling from 8% to just 1.7% in some benchmarks.

Highlights of Stanford HAI's 2025 Artificial Intelligence Index Report

Furthermore, the cost of inference for systems reaching the level of GPT-3.5 dropped by more than 280 times from November 2022 to October 2024. At the hardware level, costs dropped by 30% per year, while energy efficiency improved by 40% per year.

The threshold for advanced AI is rapidly decreasing. Not to mention the development of sparse models like DeepSeek, where only relevant parameters are activated to answer the user’s query under the Mixture of Experts (MoE) structure, making the whole thing more efficient.

Indeed, as smaller but more powerful AI models continue to emerge, the requirements for AI model training have been reduced, and cost-effective distributed training is expected to become mainstream in the next decade. There are currently some top projects conducting related research based on different theoretical frameworks.

3. AI is increasingly integrated into everyday life

In 2023, the U.S. Food and Drug Administration (FDA) approved 223 AI-assisted medical devices, up from just six in 2015. On the roads, self-driving cars are no longer experiments: Waymo, one of the largest operators in the U.S., provides more than 150,000 self-driving rides per week, and Baidu’s Apollo Go fleet of driverless taxis is now operational in several Chinese cities.

Highlights of Stanford HAI's 2025 Artificial Intelligence Index Report

4. Corporate investment in AI has increased significantly, driving record investments and adoption

The adoption of AI in business is also accelerating: 78% of organizations are using AI in 2024, up from 55% the year before. At the same time, a growing body of research confirms that AI can increase productivity and help close skills gaps across the workforce.

In fact, product-market fit breakdowns will occur more frequently as AI causes customer expectations to grow exponentially, making existing solutions obsolete overnight, leaving incumbents with no chance to adapt.

5. Despite rising optimism about AI globally, Asians are more optimistic about AI

In countries such as China (83%), Indonesia (80%) and Thailand (77%), the majority believe that the benefits of AI products and services outweigh the risks. In contrast, optimism remains far lower in places such as Canada (40%), the United States (39%) and the Netherlands (36%).

However, this attitude is changing: since 2022, optimism has grown significantly in a number of previously skeptical countries, including Germany (up 10%), France (up 10%), Canada (up 8%), the United Kingdom (up 8%) and the United States (up 4%).

Highlights of Stanford HAI's 2025 Artificial Intelligence Index Report

6. The influence of artificial intelligence in scientific research is increasing, and it has become an important driving force for scientific progress

The growing importance of AI is reflected in major scientific prizes: two Nobel Prizes were awarded for research into deep learning (physics) and its application to protein folding (chemistry), while the Turing Award recognized groundbreaking contributions to reinforcement learning.

Clearly, AI is advancing at an exponential and unexpected pace, which is significant to most people. As a result, AI security is becoming increasingly important. While AI makes forgery easier, cryptography makes it more difficult. Look forward to crypto projects that can leverage the native properties of blockchain (verifiability and transparency) to build practical solutions in this area.

Related reading: Interview with Chris Dixon, founder of a16z: The intersection of artificial intelligence and encryption technology

Disclaimer: The articles reposted on this site are sourced from public platforms and are provided for informational purposes only. They do not necessarily reflect the views of MEXC. All rights remain with the original authors. If you believe any content infringes on third-party rights, please contact service@support.mexc.com for removal. MEXC makes no guarantees regarding the accuracy, completeness, or timeliness of the content and is not responsible for any actions taken based on the information provided. The content does not constitute financial, legal, or other professional advice, nor should it be considered a recommendation or endorsement by MEXC.
Share Insights

You May Also Like

Midnight Redefines Blockchain Privacy With Zero-Knowledge and Rational Design

Midnight Redefines Blockchain Privacy With Zero-Knowledge and Rational Design

As blockchain technology matures, the industry faces a critical challenge: balancing transparency with privacy. Most public blockchains expose all transaction data, creating risks for enterprises and individuals. This tension has sparked debate about whether decentralization must sacrifice confidentiality. Fahmi Syed, President of the Midnight Foundation, believes there’s a better path forward. During Token2049 Singapore at the main event venue, he outlined to BeInCrypto Midnight’s vision for “rational privacy.” Midnight’s approach uses zero-knowledge-proofs-based smart contracts to unlock selective disclosure: the ability to control what you share, when, and with whom. Please briefly explain Midnight Network and how it differs from other privacy-focused blockchains. Midnight is a new layer one blockchain built on advancements in zero-knowledge proofs. We’ve built a dual-state, public-and-private ledger architecture that enables applications to validate sensitive data using cryptographic proofs.  Through zero-knowledge proofs and purpose-built smart contract disclosure mechanisms, individuals, corporations, and machines can decide what they share, when they share it, and with whom they share it. This is what we call “rational privacy”—selective, programmable privacy that protects sensitive data by default while still enabling compliance and auditability when required. Today, most public blockchains are transparent or pseudo-anonymous, but pseudo-anonymity is not privacy – over time, identities and wallets can be exposed, tracked, or compromised. How does your approach differ from previous attempts to add privacy to public blockchains? Public ZK chains started with the likes of Monero and Zcash. These privacy-focused networks showcased how zero-knowledge proofs could protect sensitive data, but because their tokens acted as stores of value, they raised compliance concerns for not only regulators, but corporates who must adhere to KYC/KYB procedures. The next evolution was the rise in ZK rollups or ZK chains, which primarily aimed to scale blockchain transactions and later incorporated some privacy features. But when you try to retrofit privacy, there’s always a risk of exposure. At Midnight, we’ve baked privacy into the core of the network, giving you the ability to protect sensitive data and metadata while remaining auditable on-chain. This essentially enables you to build technology and applications that preserve privacy without sacrificing compliance. What is Midnight’s mechanism that enables both privacy and compliance? Private data shouldn’t sit on a blockchain. The most valuable use of private data is when value can be derived while the underlying information remains under the owner’s exclusive control. One way this can happen is via proofs and attestations. For example, proofs of identity, ownership, or accreditation. These proofs function like keys that gateway your access into deeper levels of a product, service or network.  Today, valuable data sits in silos, completely underutilised. What Midnight can do is bring such silos together to unlock shared value, without risk of exposure. Instead of sharing raw data across networks, you can provide attestations, or proofs that enable untrusted parties to operate together in a trusted manner. In this way, I see Midnight as a truth layer, through our smart contracts, you can allow disclosures or enable different parties to validate information without risk.  With Midnight, you get to choose what, when, and with whom you’re disclosing information to. People often think of privacy as trying to obfuscate or shield. We believe privacy is a starting place for compliance. Privacy with selective disclosure will enable better compliance. Midnight uses a dual-component tokenomics system with NIGHT and DUST. What motivated this design choice, and how does it address the economic challenges facing other Layer-1 blockchains? The economic model today for most blockchains is not only confusing, it’s broken. For example, you may have a Samsung phone, but you don’t pay for your Samsung phone with your Samsung shares. Why? Because your shares are an investment, your phone is just a product that you use, or “consume”. Today in Ethereum, Cardano, Solana, and other L1s, the tokens you select for investment purposes are the same assets you use to pay for transaction fees or “gas”. This is counterintuitive – for example, what happens when the token price goes up in value? Transaction costs increase, especially during periods of network congestion, meaning you’re cannibalizing your investment just to make a transaction, essentially grinding the network to a halt.  At Midnight, we’ve separated ownership and utility from consumption. NIGHT is our native utility token that gives you ownership and governance of Midnight. NIGHT generates DUST, which is a renewable, shielded resource. DUST does not function as a store of value, as it decays within seven days. Instead of paying for transactions with NIGHT, you pay with DUST, and if you own NIGHT, your supply of DUST will continue to replenish. This model ensures you’re not cannibalising your primary asset just to pay to use the network. The Glacier Drop has attracted significant attention in the community. Can you share its main objectives and how it supports Midnight’s vision? We’re so confident in our technology and its capabilities that we’re giving away 100 percent of the token supply of NIGHT through a multi-phase distribution process, starting with the Glacier Drop, which is open to users from across eight major blockchain ecosystems. If you were holding at least $100 worth of BTC, ETH, ADA, SOL, AVAX, BNB, XRP, or BAT tokens in a self-custody wallet on the snapshot date, you are eligible to come and claim. The amount of NIGHT you can claim corresponds to your ownership in the other eligible chains. The more you have there, the more NIGHT you’ll receive. Participants from each of these ecosystems have the opportunity to come and claim before we open this up to anyone during the Scavenger Mine phase.  Scavenger Mine allows anybody from any ecosystem or walk of life to claim a portion of the unclaimed tokens from Glacier Drop. Only after the end of Scavenger Mine is there a distribution to the Midnight Foundation, the on-chain treasury, and the on-chain reserves. You recently announced a collaboration with Google Cloud. How does this partnership advance Midnight’s enterprise adoption goals, and what does it mean for bringing traditional Web2 companies into the blockchain space? That’s right, our collaboration with Google Cloud is bringing enterprise-grade infrastructure support to our network, which will give institutions and others more confidence to utilize Midnight’s privacy-enhancing infrastructure. Through this partnership and others, millions of users and thousands of corporate clients are welcome to utilize Midnight’s technology to bring enhanced privacy functionality to their products and services. Can you elaborate on this partnership with a real-world example? A healthcare company in Turkey with three million patients is currently working with us to explore how they can leverage blockchain infrastructure to generate proofs of their patients’ medical histories. Our strategy is to start with partners with slightly lower regulatory hurdles for proof of concept. Once we can demonstrate viability in one area, we can extend it to another. For example, now we’re in conversations with a large hospital in California that’s looking to use Midnight for cross-clinical trials with other external partners. They want to protect sensitive patient data, so they’re looking at how Midnight can bring together different silos of medical history and records to achieve better outcomes for their patients and the medical industry as a whole, without ever exposing the data on the chain. Can you walk us through Midnight’s roadmap from testnet to mainnet launch? What are the primary milestones and goals for the rest of 2025 and beyond? Our primary goal for this year is to complete Glacier Drop successfully, launch our token, as well as preparing for mainnet launch. From there, our focus will be on how we bring our technology to market while still maintaining our path to decentralization. To build institutional confidence, our strategy is to launch with a consortium of federated nodes, made up of ten trusted partners running validators, to provide the robustness, speed, and scalability that’s needed for enterprises to operate securely and confidently.  As we scale, through feature releases, upgrades, and with partners bringing in more transactional volume, Midnight will gradually develop into a decentralized ecosystem. To support this, when we launch on mainnet, we’ll be running an incentivized testnet in parallel to the federated mainnet. Eventually, the two will converge, and we will end up with a fully decentralized blockchain where validation is not just from trusted partners, but from a wider group of 100 to 200 validators.
Share
Coinstats2025/10/09 11:00
Share