The AI Index is an independent initiative at the Stanford Institute for Human-Centered Artificial Intelligence (HAI), conducted by the AI Index Steering Committee, composed of an interdisciplinary group of experts from academia and industry. The annual report monitors, examines, extracts, and visualizes data related to artificial intelligence, supporting decision-makers to take impactful action to advance AI responsibly and ethically with humans in mind.
The latest edition of the Stanford AI Index report incorporates data from a broad set of academic, private, and nonprofit organizations, self-collected data, and original analysis than any previous editions. The 2022 issue includes an expanded technical performance chapter, a new survey of robotics researchers from around the world, data on global AI legislation records in 25 countries, and a new chapter dedicated to examining technical AI ethics metrics.
Humans.ai is committed to its goal of creating a unique infrastructure tailored to meet the evolving demands of artificial intelligence research and development. The approach envisioned by Humans is to bring AI to a Blockchain infrastructure to democratize this technology, increase its security and streamline its development.
Concerning the findings of the Stanford report, Humans.ai is already breaking new ground in the field by complying with key elements like facilitating the ethical development and use of this technology, researching and developing new ways to use voice models and democratizing access to AI.
Stanford AI Index 2022 - study overview
The 2022 AI Index report, released by the Stanford Institute for Human-Centered Artificial Intelligence (HAI), provides a comprehensive overview of the state of artificial intelligence (AI) and its impact on society. The report covers a wide range of topics, including research and development, industry trends, and societal implications of AI.
One of the report's key findings is that AI research continues to grow rapidly. In 2020, the number of AI-related publications in academic journals reached an all-time high, and the number of AI-related patents filed also increased. These developments indicate that more and more researchers and companies are investing in AI and that the field is advancing quickly.
The report also looks at the impact of AI on various industries. It notes that the use of AI in healthcare has been particularly promising, with applications such as image analysis and natural language processing helping to improve diagnosis and treatment. In addition, AI has been implemented in a series of key areas like finance, transportation, and retail, where it boosts efficiency and improves decision-making.
Focus is also placed on the societal implications of AI, the authors noting that AI can have a positive impact by helping to solve complex problems and improve people's lives. However, the paper also highlights the potential risks, such as the potential for job displacement and the need for ethical considerations when developing and deploying AI.
A closer look is also taken at the geographical distribution of AI research and development. The report underlines that the United States, China and Europe continue to lead the AI research race, but other regions, such as Asia, Latin America, and Africa, are also steadily closing the gap. The report also found that collaboration between countries is becoming increasingly common, with researchers from different countries co-authoring papers and working on projects.
One of the more interesting sections of the report analyses the diversity levels within the AI community. Women and underrepresented minorities seem to continue to be underrepresented in the field of AI, both in terms of the percentage of researchers and the number of papers published. The report also signals some disparities in the representation of different ethnicities and races in the field, underscoring the fact that more efforts and initiatives are needed to stimulate diversity and inclusion in AI, pointing out that diverse perspectives can play an important role in ensuring that AI is developed ethically and fairly.
The report also addresses the popular fear concerning how AI will impact the job market, pointing out that this technology will inevitably lead to a level of job displacement through the automation it provides. Even so, AI will generate new jobs, such as data scientists and AI engineers. The bottom line is that AI will not remove the human element from the job market. Instead, it will become a new instrument that will work alongside people to streamline their activities and boost productivity while eliminating repetitive tasks.
Therefore, the report emphasized the importance of retraining workers for emerging jobs and providing support for those seeking a professional reconversion.
Overall, the 2022 AI Index report provides a valuable snapshot of the current state of AI and its impact on society. It shows that the field of AI is growing rapidly and that it has the potential to bring many benefits to society. However, it also highlights the need for careful consideration of the potential risks and the importance of taking a human-centered approach to the development and use of AI.
In conclusion, the 2022 AI Index report provides a comprehensive and in-depth analysis of the state of artificial intelligence. It illustrates how AI is impacting different industries and society as a whole. The report highlights the need for continued research and development in AI and the importance of addressing ethical concerns, diversity and inclusion and job market implications. The report serves as a valuable resource for policymakers, researchers, industry leaders, and anyone interested in the field of AI.
Key takeaways related to Humans.ai’s approach to AI
Private investment in AI soared while investment concentration intensified
Private investment in AI has skyrocketed in the past two years. The report highlights that private investment in AI in 2021 reached a staggering $93.5 billion, more than doubling the total private investment in 2020. In contrast, the number of newly funded AI companies followed the same descending trend of the past couple of years, from 1051 companies in 2019 and 762 companies in 2020 to 746 companies in 2021. The number of funding rounds has almost quadrupled, from 4 funding rounds worth $500 million or more in 2020 to 15 funding rounds in 2021.
Even if AI is garnering more attention than ever from investors, the process of investing in this technology is convoluted and, to a certain degree, inaccessible to most people. Humans believes that AI belongs to humanity as a whole and that only through our combined efforts we can create ethical AI.
Humans.ai has taken it upon itself to build a multifaceted ecosystem that addresses some of the major dysfunctions present in the current AI status quo. The solution proposed by Humans.ai fuses artificial intelligence and blockchain, a technology that will play a pivotal role in the upcoming Web3 revolution, to set the foundation for an ecosystem that will unite AI developers, researchers and consumers. As AI is slowly consuming the world, enriching industries and businesses, Humans.ai has made it its mission to assist the transition of artificial intelligence, AI models and AI-based solutions from the centralized Web2 domain, where development is done behind closed doors where giants of the tech industry hold the reigns, to a decentralized paradigm, where AI becomes a democratized asset that is constantly monitored by real humans to ensure its ethical use.
The Humans ecosystem answers the question raised by the bottlenecks present in the current AI investment process by reuniting VCs and people, regardless of their technical background, around the common goal of financing the development of the AI of the future and paving the way towards the AI economy. Scale, a component of the Humans Ecosystem, promotes an alternative way of investing in AI based on decentralization and collaboration provided by blockchain technology. By resting at the intersection of AI and blockchain, Scale creates an economy around each AI product to which people can openly and transparently contribute.
Language models are more cable than ever, but also more biased
Technological advancements and access to increasingly large data pools for training have enabled researchers to create large language models that set new records on technical benchmarks. But even with their technical complexity, large language models signal a worrying susceptibility to picking up biases from their training data. The survey underlines that a 280 billion parameter model developed in 2021 exhibits a 29% increase in elicited toxicity compared to a 117 million parameter model considered state of the art in 2018. Undoubtedly, language models are growing more capable over time, but so does their propensity for biases.
Humans.ai has created working extensively on a library of unique synthetic voices, each of them with different pitches, tonalities and ways of expressing themselves. Compared to other similar voice model projects, Humans.ai wishes to expand the concept, giving it a wider range of possible applications by focusing on the transition from one modality to another, namely text to speech. As such, the Humans AI Voices can take text input and use the to create an audio output. The technology developed by Humans does this by translating the text and language to phonemes, the smallest unit of sound that distinguishes one word from another word in a language. Based on these phonemes, the AI developed by Humans.ai performs the transition from text to audio.
The Voice of Community is a unique use case that leverages Humans.ai’s synthetic voice technology which combines multiple voices donated by community members using state-of-the-art AI technology like Natural Language Processing and Machine Learning to create a unique voice to become the voice of a community or a brand ambassador.
Talkens.ai is an AI experiment launched on Humans.ai’s Launchpad. Talkens.ai empowers NFT holders by allowing them to take their digital collectables to the next level, turning them from a standard jpeg image that resides on top of the blockchain into a smart asset class capable of speaking in a unique voice.
The Talkens AI NFT is a new twist on the non-fungible token concept. Each Talkens AI NFT is represented by a unique design that has been generated with the help of artificial intelligence. What makes them truly unique, besides the awesome design, is the fact that each of them encapsulates a unique synthetic voice that can be used to bring regular NFTs to life.
The rise of AI ethics everywhere
Since 2014, ensuring fairness and transparency in AI has dominated the minds of researchers all over the world, a fact demonstrated by the fivefold increase in related publications at ethics-related conferences. What is game-changing is that algorithmic fairness and bias are no longer confined to the academic sphere, embedding themselves as a prevalent research topic having deep implications across a wide range of fields. The Stanford report underlines that researchers with industry affiliations submitted 71% more publications at ethics-focused conferences in the past years.
Humans.ai is creating a Web 3.0 blockchain infrastructure governed by a consensus mechanism in which real humans play a crucial role as validators in a framework designed to keep AI honest. This new consensus mechanism is called Proof of Human, and it relies on humans to leverage their biometric data to prove that a specific AI is still under close biological supervision.
The aptly named Humans Blockchain allows creators to own all their digital assets. The development of an all-in-one platform for AI will offer unprecedented creative possibilities, and blockchain technology is the solution to control this environment, the ultimate goal being to comply with AI ethics.
AI is becoming more affordable and higher performing
Training image classification systems has decreased by a staggering 63.6% since 2018, while the time required for training has improved by a whopping 94.4%. The decrease in AI training costs and the significantly reduced training times have spread across multiple Machine Learning areas such as recommendation, object detection and language processing, favouring a better commercial adoption of AI technologies.
Humans.ai intends to create an economy around each AI that will live inside its ecosystem, making it more affordable and simpler for anyone to contribute to the development of artificial intelligence. Furthermore, by employing blockchain as a foundation for its infrastructure, Humans.ai decentralizes AI, making it more secure and private.
Decentralized AI distributes processing across a network of devices, enabling faster and more efficient processing, making it invaluable for applications that require real-time decision-making, such as autonomous vehicles. In contrast, in centralized AI systems, all processing takes place in a single location, which usually leads to bottlenecks and traffic congestion that result in slower response times.
Other major takeaways from the Stanford study
Data, data, data
Unsurprisingly, reliance on extra training data seems to be the key to generating better results across technical benchmarks. Statistics show that as of 2021, 9 state-of-the-art AI systems out of the 10 benchmarks analysed in the Stanford report were trained with extra data. This reliance on additional datasets favours private sector players, who have access to vast datasets.
More global legislation on AI than ever
The report highlights that an AI Index analysis of legislative records on AI in 25 countries confirms that the number of bills containing the words "artificial intelligence" passed into law went up, from just one entry in 2016 to a total of 18 in 2021. The countries with the highest number of AI-related bills passed in 2021 are the United States, the United Kingdom and Spain, each adopting three such bills.
Robotic arms are becoming cheaper
The average price of robotic arms has decreased by 46.2% in the past five years, from an average of $42,000 per arm in 2017 to $22,600 in 2021, which highlights that robotics research and development have become more accessible and affordable.
