This article offers you a detailed commentary on AI, with numerous current examples from the Anglo-Saxon press. It’s ideal for revising before a glue or oral exam!
The transition after the summary
Transition : In 2023, the Dr Hintonwhich created the technology that became the foundation of AI systems, has joined a group warning about the dangers of AI. He later left Google, saying these companies will exacerbate the AI threat.
Problematic : Therefore, we can ask ourselves to what extent the development of AI represents a threat to humanity ?
Advances in AI actively contribute to the advancement of society
AI enables many advances in medicine
Baseimmunea British biotech startup, is using AI to help it predict future mutations in pathogens, with the aim of creating coronavirus malaria vaccines that will be more effective and last longer. Baseimmune thus obtained seed funding from $14 million from MSD Global Health Innovation Fund to develop these vaccines “stands the test of time » thanks to artificial intelligence.
The Golborne Medical Centerlocated in one of the most deprived areas of the UK, has revealed the creation of a stethoscope using AI. This stethoscope, Eko AIwas designed to detect heart disease in seconds. It will revolutionize diagnostics and dramatically reduce NHS waiting times, experts say.
In a study published in May 2024 in the journal Drug Discoveryfive experts showed that The world’s top 20 pharmaceutical groups are now all using AI to accelerate the discovery of new molecules. 67 clinical trials were launched in 2023 (10 times more than five years ago).
AI promotes the development of innovations
Since February 2022, with the presentation of FourCastNet pair NvidiaAI has transformed the management of natural disasters improving the speed and accuracy of weather forecasts. Unlike traditional models that require two to three hours on supercomputers for a simulation, AI systems like FourCastNet produce results in just one minute, while providing comparable accuracy. This advancement enables faster and more accurate alerts, facilitating more effective preparation and response to natural disasters, thereby reducing risks and potential damage.
According to McKinseyl’Generative AI is on the verge of revolutionizing the world of work by significantly boosting productivity thanks to its innovations. By providing automated tools to create and optimize processes, AI could increase work productivity for 0.1 to 0.6% per year until 2040. This advancement promises continued improvement in efficiency across various economic sectors, enabling businesses to achieve substantial productivity gains while boosting economic growth.
AI can revolutionize agriculture
For the “ AI Institute for Resilient Agriculture “, the use of AI could enable the development of a strategy called ” precision agriculture » which could optimize production so that crops grow in favorable conditions. The idea is that farmers use AI-enabled technologies (for example: drones that assess a field’s water or pesticide levels) to precisely target problem areas.
According to the director of government affairs of theAmerican Farm Bureau Federationprecision agriculture could reduce water waste and increase productivity, making it possible to produce as much as 30 years ago with 100 million hectares less.
However, spectacular advances in AI pose a risk to our societies
The rise of deepfakes constitutes a threat to democracy
AI has enabled the emergence of deepfakesthat is to say photos, videos or audios made or modified using AI. These deepfakes jeopardize the democratic process, as they can sow confusion among citizens/voters, stimulate disinformation, discredit political figures and unduly influence the outcome of elections.
A striking example of this threat emerged in October 2023, when deepfakes audio made it appear that Keir Starmer was insulting his staff during the first day of the Labor Party conference in October 2023. Although fact-checkers determined the audio was false, it surpassed 1.5 million hits on X and damaged Labor’s image during this conference.
THE deepfakes Audio can also be used to sow discord and incite violence. In February 2024, a deepfake Audio of London Mayor Sadiq Khan disparaging Remembrance Weekend and calling for pro-Palestine marches in London was made to look like a secret audio recording. This recording went viral on social media and sparked hateful comments against him.
The environmental damage generated by AI is colossal
The International Energy Agency estimates that the amount of energy ChatGPT uses to meet each request is 2.9 watt hours (Wh), 10 times more than a Google search.
Cooling data centers are also giving AI chatbots an incredible thirst. In 2023, another study from the American University of Colorado calculated that asking 25 questions to ChatGPT cost half a liter of water. Enough to reach millions of liters if you multiply that by its 200 million users.
The challenge posed by AI is clear. According to International Energy Agency (IEA), data center demand will almost double by 2026, from 460 TWh to 1,000 TWh. The risk is that the data center boom will spur the construction of new gas-fired power plants. For example, Microsoft is building a gas-fired power plant in Dublin to provide backup power.
AI also threatens some jobs
In its report published in January 2024 (“ Gen-AI: Artificial Intelligence and the Future of Work ), the International Monetary Fund is concerned about the potential impact of artificial intelligence on employment. The organization suggests that almost 40% of jobs worldwide will be affected by AI technologies and that around 60% of these jobs will be in high-income countries.
The IMF warns that about half of them could benefit from increased productivity, which would exacerbate social inequality around the world.
The growing need to regulate AI for global security
A step towards global governance: AI Safety Summit
The AI Security Summit (AI Safety Summit), organized by the British government on 1is and November 2 at Bletchley Park, marks a significant step forward, as it is the first time that global decision-makers have come together to seriously discuss AI.
This summit reflects a awareness growing awareness of the challenges and opportunities presented by artificial intelligence, highlighting the importance of international regulations and collaborations to ensure that AI developments are safe and beneficial for society.
The pioneering regulation of the European Union
The AI regulation (AI Act) of the European Union provides a crucial regulatory framework for assessing the risks posed by AI technologies. This framework includes four levels of risk: unacceptable, high, limited and minimal. The importance of this regulation lies in its ability to classify AI technologies according to their potential danger to the health, safety and fundamental rights of individuals.
At the same time, the EU or initiatives like McAfee’s Mockingbird project aimed at detecting and exposing audio alterations in videos, show concrete efforts to combat deepfakes. Companies providing AI voice generation services are also taking steps to ensure their systems can identify altered audio, contributing to proactive monitoring of these technologies.
The UK AI Bill
The UK AI billexpected later this year, will implement existing voluntary agreements between businesses and government by making them legally binding. The Labor Party, under the leadership of Keir Starmer, plans to introduce formal rules for AI, as outlined in its election manifesto.
Starmer’s government has pledged to introduce binding regulations for companies developing the most powerful AI models and to legislate to ban deepfakes sexually explicit. Targeting the most powerful AI models, Labor would impose tighter restrictions on companies such as OpenAI, Microsoft, Google, Amazon and AI start-ups, such as Anthropic, Cohere and Mistral.
Conclusion
Overall, as the world transitions fromNarrow AI to artificial general intelligencehe becomes aware of the double-edged sword that it represents. The dangers must not be underestimated and the regulations must be structured around two axes, namely the transparency et access. The now famous software Pegasus is a prime example of the potential misuse of technology. This does not have to be the case for artificial intelligence.
If you would like to continue reading on artificial intelligencego to this article or this one!
Related News :