show index hide index
Artificial intelligence arouses as much fascination as concern about its potential impact on the future of humanity. A researcher specializing in AI has just revealed an alarming figure that questions: could AI really cause the end of the world? A serious question that deserves our full attention and in-depth reflection.
Worrying predictions are emerging around the rapid evolution of technologyartificial intelligence (AI). Renowned specialists are increasingly expressing their fears about the potential dangers of AI for humanity. One of these researchers, Roman Yampolskiy, professor of computer science at the University of Louisville, says that the rise of AI could lead to human extinction with almost certain probability.
Extinction probabilities: what are the predictions?
Roman Yampolskiy shared his concerns during a podcast hosted by Lex Fridman. He believes that the advent ofAGI (general superintelligences) has a 99.9% chance of causing the annihilation of humanity within the next 100 years. This alarmist vision is based on the ability of AI to make billions of decisions per second, which could be beyond our understanding and control.
Concrete dangers: examples and diversions
To illustrate his fears, Yampolskiy refers to several cases ofaccidents and misuse of AI technologies. For example, the “God fashion” from GPT-3, recently blocked by OpenAI, shows that even the most secure systems can be manipulated. These diversions highlight the potential flaws of these technologies and the inability of creators to foresee all possible situations.
Diversity of opinions
Although Roman Yampolskiy is particularly pessimistic, he is not the only one to warn of the risks of AI. Yann LeCun, head of AI at Meta, also shares similar concerns, although the scientific community is not unanimous on this topic. Many experts point out that the science of predicting the impacts of AI is still in its early stages and could progress in positive ways.
The opinion of former OpenAI employees
A collective of former and current OpenAI employees recently wrote an open letter criticizing the company’s lack of caution in developing AI. According to Daniel Kokotajlo, former researcher at OpenAI, the company could finalize a AGI by 2027, with potentially disastrous consequences for humanity. The collective calls for increased security measures and more rigorous attention to the risks posed by these technologies.
Recommendations for a secure future
- Setting up safeguards for each new technological development.
- Strengthening regulatory and ethical frameworks around AI.
- Development of international collaborations for global risk monitoring and management.
An evolving science
It is important to note that, despite these highly anxiety-inducing predictions, the research and science around AI continues to evolve. Advances in security, ethics and transparency could fortunately mitigate current risks. At the same time, other existential threats such as pandemics and nuclear weapons remain major concerns for our collective present and future.
To read LinkedIn : le grand ménage débute, place aux posts authentiques sans IA
In conclusion, although the rise ofartificial intelligence presents tangible risks, it is paramount that we continue to monitor and regulate these technologies to ensure that they serve humanity in a beneficial and safe way.
