Qual è il piu grande rischio dell’AI? | Filomena Floriana Ferrara | TEDxLink Campus University

Key Concepts

• Intelligenza Artificiale (AI): A discipline that studies systems designed to simulate human cognitive abilities with the goal of improving human quality of life.
• Storia dell'AI: Traced from early theoretical concepts to its formal inception in 1956 and subsequent developments.
• Machine Learning: A subset of AI that enables systems to learn from experience without explicit programming.
• Reti Neurali Artificiali: Computational models inspired by the structure and function of biological neural networks.
• AI Act: European Union legislation aimed at regulating AI to ensure it is responsible, transparent, secure, and ethical.
• Alfabetizzazione Digitale: The ability to interact with technology and digital systems, now considered a fundamental skill by UNESCO.
• Innovazione e Lavoro: The historical pattern of technological innovation leading to job displacement in some sectors but creating more jobs overall.

The Greatest Risk of Artificial Intelligence

Floriana Ferrara, Corporate Social Responsibility Manager at IBM Italia, inventor with 21 patents, and dyslexic, addresses the question: "What is the greatest risk of artificial intelligence?" She argues that the greatest risk is not using AI.

Origins and Evolution of Artificial Intelligence

Ferrara begins by tracing the history of AI, emphasizing that it is not a new concept.

• Early Concepts: Before 1900, AI was primarily theoretical, explored by writers who imagined thinking machines. Jonathan Swift's "Gulliver's Travels" is cited as an early example, featuring a machine that generated ideas and phrases.
• Shift to Concreteness (Early 1900s): The transition from fantasy to reality began around 1900. The definition of AI is presented as a discipline that simulates human mental capabilities to improve human quality of life. This purpose, Ferrara notes, has been consistent across all major innovations, from trains to energy.
• The Calculator (Early 1900s): The calculator is presented as the first concrete example of AI, simulating the human ability to perform calculations to improve the lives of accountants and individuals.
• Early AI Milestones:
  • 1914: Spanish engineer Torres creates a simple simulation of a chess-playing machine.
  • 1950: Alan Turing publishes his seminal paper "Can Machines Think?", laying the groundwork for AI research.
  • Mid-1950s: Marvin Minsky, a Harvard professor, develops the first artificial neural network and a simulator for mouse behavior in mazes.
• The Birth of the Term "Artificial Intelligence" (1956): Four prominent scientists—John McCarthy (Dartmouth College), Marvin Minsky (Harvard), Nathaniel Rochester (IBM), and Claude Shannon (Bell Company)—organized a workshop. Their goal was to survey existing projects related to "thinking machines" and to give the discipline a formal name. They coined the term "Artificial Intelligence" and celebrated its "birthday" on August 31, 1956.
• Early Programming and Machine Learning:
  • 1958: John McCarthy invents LISP, the first programming language for AI.
  • 1959: Arthur Samuel develops the first machine learning program, a checkers-playing software that learned from its games.

Ferrara highlights that these foundational elements of AI, such as machine learning and neural networks, are not recent developments but have been studied for decades.

AI's Journey to Public Awareness

Ferrara then discusses key moments that brought AI into the public consciousness.

• IBM's Deep Blue vs. Kasparov (1996-1997): IBM's AI, Deep Blue, defeated world chess champion Garry Kasparov. This event generated significant media attention, making the public more aware of AI's capabilities.
• IBM's Watson on Jeopardy! (Early 2000s): IBM's AI, Watson, competed on the popular American quiz show "Jeopardy!" against two human champions and won. This further demonstrated AI's advanced capabilities in natural language processing and knowledge retrieval, creating another "wow" moment for the public.

Understanding AI: Not a Monolithic Entity

Ferrara cautions against viewing AI as a single, monolithic entity.

• Diverse Systems: She emphasizes that AI is not a single system that accumulates information but rather comprises numerous diverse systems developed by different companies, research centers, and universities.

The Impact of Generative AI and the AI Act

The advent of generative AI, particularly ChatGPT, marked a significant turning point.

• ChatGPT's Arrival (November 30, 2022): The widespread availability and use of ChatGPT brought AI into homes, businesses, and schools, making it accessible to everyone.
• Responsible Use: While acknowledging its utility, Ferrara stresses the importance of not delegating personal preparation and critical thinking to AI. She reiterates the core definition of AI: simulating human capabilities to improve quality of life.
• The AI Act (2024): The European Union's AI Act is presented as a crucial regulatory framework. It is the first law of its kind globally, designed to ensure that AI development and use are responsible, transparent, secure, and, most importantly, ethical. Ferrara states that all companies, researchers, and users desire ethical AI.

The Need for Interdisciplinary Collaboration and Education

Ferrara underscores the evolving nature of AI and the necessity of broad understanding and collaboration.

• UNESCO's Perspective on Literacy: Citing UNESCO, Ferrara explains that just as basic literacy in the past required reading, writing, and arithmetic, modern literacy (from the 2020s onwards) now includes interacting with technology and digital systems, particularly AI.
• AI as an Interdisciplinary Field: For the first time, a technological discipline requires the integration of all other disciplines to evolve and improve. Ferrara, as a computer scientist, acknowledges her limitations in fields like medicine or law and the need for AI to be informed by experts in these areas to be truly prepared, ethical, and responsible.

Historical Parallels: Innovation and Job Creation

Ferrara draws parallels between the fear surrounding AI and historical reactions to technological advancements.

• The Calculator: Even the creation of the calculator, a seemingly simple tool, was met with fear and resistance, with concerns about job displacement.
• The Train: The introduction of trains, which revolutionized travel, was also met with apprehension about replacing jobs previously held by those involved in horse-drawn carriages. However, Ferrara argues that while some jobs were lost, the train created a significantly larger number of new jobs (e.g., track construction, station staff, ticket agents).
• General Principle: She concludes that every innovation, while potentially displacing some jobs, ultimately leads to the creation of a greater number of new employment opportunities.

The Greatest Risk: Non-Utilization

Ferrara circles back to her initial question.

• The Core Argument: After detailing AI's history, its evolution, its current regulatory landscape, and its potential for positive impact, Ferrara asserts that the greatest risk of artificial intelligence is not using it. This implies a missed opportunity to leverage AI for human betterment and progress.