Il ChatGpt success and new applications of artificial intelligence has made this new technology hit the headlines. But say what we talk about with artificial intelligence? Taking as definition that of the European Parliament for artificial intelligence of course the ability of a machine to display human capabilities such as reasoning, learning, planning and creativity. Artificial intelligence allows systems to understand their environment, relate to what they perceive and solve problems, and act towards a specific goal. The computer receives the data (prepared or collected by sensors, such as a video camera), processes it and responds.
to understand theevolution of this technology comes to our aid theStanford University which, through his Human-Centered Artificial Intelligence (HAI), no track growth During the years. And, according to the University, 2021 was the year of the turning point: “The 2021 it was the year that AI transitioned from one emerging technology to mature technology: it is no longer a speculative part of scientific research, but something that has real-world impact, both positive and negative,” he commented. Jack Clark, co-president of theAI Index.
What is the AI Index Report?
THEAI Index Report it's a annual report compiled by the Institute for Human-Centered Artificial Intelligence (HAI) at Stanford University. Arrived at fifth edition, I study analyze the main trends and the state of the art on artificial intelligence. Data is tracked, collected and analyzed as the AI advances.
The report is written by experts from academia and industry. It aims to analyze AI with a human-centered approach and the consequences of its ethical impact on society. The topics covered cover all areas related to technology such as research and development, technical performance or product laws.
The key elements of the AI Index Report 2022
“This year's report tells us that theAI is integrated into the economy and its effects are starting to go global through research, implementation and even funding,” he explains Jack Clark. A few can be identified from the report key points:
In 2021, the private investments they reached the figure record of 93,5 billion dollars, equal to double the total investments in 2020. On the other hand, the number of new companies receiving financing is decreasing: from 1051 in 2019 to only 746 in 2021. The growth in the value of individual investments is increasing: if in 2020 there were 4 funding rounds worth $500 million or more, there were 2021 in 15.
An interesting fact that emerges from the report is that concerning the close cooperation in the sector between United States and China. Despite the growing geopolitical tensions, in fact, from 2010 to 2021 the United States and China have collaborated a lot on topics relating to theartificial intelligence proving to be the major countries to collaborate in the sector. Suffice it to say that the second highest cooperation, that between the United Kingdom and China, produced around 2,7 times fewer publications.
I linguistic models are more and more capable but also conditioned by prejudices with which they are trained. In fact, large linguistic models (LLM) and deep learning algorithms (Deep Learning) despite being trained on huge amounts of data, however, still prove to be unable to overcome the "bias" with which they are trained. In practice, over time, systems are becoming more and more efficient, but as their capabilities increase, so does the potential severity of their biases.
Also increases the need for data. As of 2021, 9 of the 10 state-of-the-art AI systems reviewed are trained with additional data. This implicitly favors private sector actors who have access to various data assets.
Theimportance of the ethical issues of artificial intelligence. Since 2014, publications related to the topic have increased fivefold. Research on fairness and transparency in AI has exploded since 2014, with a fivefold increase in publications related to ethics conferences. Algorithmic fairness and bias have gone from being primarily an academic research to becoming firmly entrenched as a mainstream research topic with wide-ranging implications. Researchers with industry affiliations have contributed 71 percent more publications year-over-year to ethics-focused conferences in recent years. “As AI systems become more and more capable, it becomes important to measure and understand the ways in which they can perpetuate the harms", he claims Helen Ngo, researcher and co-author of the chapter dedicated to AI ethics.
Globally, with the evolution of AI, it begins to expand also there regulatory about. From 2015 to today in the legislatures of 25 Countries 18 times as many AI-related bills were passed worldwide. AI mentions have increased 7,7 times over the past six years. Only the United States has proposed 130 laws in 2021 (in 2015 it was only one). Still low, however, the bills approved with only 2% eventually becoming law in the past six years. Spain, the UK and the US passed the most AI-related bills in 2021, with three each adopted.
Since 2018 I am AI costs decreased (-63,3%) while increasing performance with 94,4% improved training times. The trend of lower costs affects categories such as speech processing and object detection. Factors thus driving the commercial adoption of AI technologies.
Also the robotic research has become more accessible and convenient. The average price of robotic arms has decreased by 46,2% over the past five years, from $42.000 per arm in 2017 to $22.600 in 2021.
The technological challenge between Microsoft and Google
La race for artificial intelligence is more alive than ever, with the giants of Big Tech, ready to challenge each other on what is considered the new frontier of technology. In particular Microsoft e Google they started to challenge each other for dominance in the field.
Microsoft he anticipated the moves of his rivals investing 10 billion in OpenAI, the parent company of ChatGPT and believed to be one of the world's top three laboratories for artificial intelligence. The investment has allowed the Redmond company to integrate the AI platform into its structure. Now, Chat GPT è usable through Bing search engine and was integrated into also Microsoft Teams. The goal of storming the hegemony (and market) of the search engine dominated by Google has begun. And the numbers are promising: globally they are downloads tenfold which involved not only the search engine application, but also the Microsoft home browser, Edge. Meanwhile Chat GPT He has reached stratospheric numbers: just two months after launch has reached 100 million monthly active users winning the title of fastest growing consumer application in history.
Google she was caught unprepared by the moves of the Redmond company. She called her pioneers back, Larry Page e Sergei Brin, and speeded up its AI program. In recent days, Google has announced an investment from 300 million dollars in the AI start-up, anthropic.
The reply to ChatGPT also arrived with theBard's introduction, its Chatbot. Bard aims to generate detailed answers to simple questions. Its operation is based on the LaMDA language, the Language Model for Dialogue Applications. Bard's presentation, however, proved to be a sensational own goal for the company: in the promotional video for the launch of the AI, Bard was completely wrong in answering to a question about the James Webb telescope. A fool which did lose in just 1 day00 billion dollars of capitalization with the shares of Alphabet, the parent company of Google, which they lost 8% on the stock market. A hard blow for the Mountain View giant.
However, the technological challenge is only just beginning. We will still see some good ones.
