Artificial Intelligence: ethics and trends

Recently, several posts have been published on LinkedIn, and searches on major browsers have increased: a mix of curiosity and avant-gardism but also hostility and fear has characterized the dialogue around Artificial Intelligence. General attention has almost exclusively focused on Google’s alleged sentient AI and generative AIs, particularly around Chat GPT. Indeed, last summer, a Google employee claimed that Google’s language model, LaMDA, was capable of giving responses influenced by love, suffering, anger, and thus had the ability to feel emotions. All these statements were immediately denied by the company, but this did not prevent ethical doubts from arising around the issue. Subsequently, OpenAI developed and made public Chat GPT, the language model based on deep learning, capable of conversing on any topic, giving articulate and well-written responses. The chatbot immediately attracted strong attention: in less than two months, it reached 100 million users, and a first paid version was released, offering more sophisticated features.

But Artificial Intelligence is not just this. Many of its valuable applications often remain unknown to most. Some examples can help understand the concrete uses of Artificial Intelligence and the benefits they bring, not only at the enterprise level.

Areas of intervention and concrete benefits of AI

• Agri-food sector: in agriculture, having the intuition of the right moment is important, but often not enough. AI can collect and process numerous data, making it possible to have a reliable and precise analysis of weather phenomena. Knowing this information helps farmers better plan the timing of sowing and resource management. Crop supervision is also favored: through computer vision, it is possible to easily check for any damage to plants and fruits and the optimal level of product ripeness. Concrete applications can go even further: some researchers at the Catholic University have trained algorithms on 408 samples of Taggiasca olive oil to prove the authenticity of the oil’s origin and avoid food fraud.

• Medical field: AI advancements in the medical field are increasingly significant. Machine Learning has been used multiple times in cardiology, sleep disorder detection, integrating IoT systems in healthcare, and preventing diseases and pandemics. Recently, an algorithm called Sphinks was developed, capable of analyzing tumor proteins and enzymes typical of malignant cells. This makes it easier to understand which family the tumor belongs to and slow down the disease’s progression. Some researchers have also developed an AI that detects atrial fibrillations, identifying the risk even in the absence of current anomalies.

• Environmental and wildlife protection: AI can help the ecosystem by leveraging the learning capabilities of algorithms. For example, some studies have allowed the interpretation of abnormal sounds emitted by chickens, detecting any anomalies in their health and farm conditions and intervening accordingly. A similar project has also been carried out with corals: through sound, algorithms can detect degraded parts of the coral reef with 92% accuracy through their own “songs”. The goal, in this case, is to identify damaged parts and optimize marine environment protection.

• Conservation and research: present and past can “shake hands”: this is demonstrated by the application of AI for the protection of archaeological heritage; a team of Italian engineers has developed an AI-based sensor system that, by adapting to weather conditions, constantly monitors historical artifacts lying on the seabed. Studies and research are progressing in parallel: according to Natural Computational Science, in the not-too-distant future, Artificial Intelligence will help astrophysics by integrating and analyzing billions of data from telescopes, antennas, and other instruments monitoring the cosmos.

The future between fears, good intentions, and regulations

Some fear that technology will end up wiping out many of today’s jobs and that the progress linked to these innovations will stifle human creativity. However, two factors must be considered.

Every technological innovation, from the invention of the World Wide Web to the launch of the first real Social Network in 1997, has caused perplexity, discontent, and simultaneous enthusiasm: some started browsing and/or chatting driven by the “serendipity” effect, while others delayed joining a social network for fear of sharing their identity through a screen.

Another important element is that each of these epochal innovations has brought with it problems to solve, dark sides to regulate, and more or less correct and ethical uses. Among them, the loss of sensitive data and hate speech. Innovative tools are valuable and can be used, like AI, in many fields with improving effects, but equally fundamental is how they are used and the regulations issued to govern them. And this does not depend on an algorithm but on human intelligence.

The first steps have already been taken. The European Union has reached a “general orientation” focused on the importance of AI use in modern society, an AI that must be safe and respect fundamental rights and the values of European citizens. The general direction aims to ban some uses and highlights others to be considered high-risk. The European agenda foresees further measures on this, which will probably not be long in coming. At the national level, the United Kingdom, for example, has planned to draft a document containing guiding principles for AI use (safety, responsibility, transparency…). Additionally, the International Conference on Machine Learning has banned the use of systems like Chat GPT for writing academic papers. The road ahead is still long, but the intentions to act are already numerous.