show index hide index
- The poorly addressed challenges of artificial intelligence
- Another example is La Banque Postale’s use of software for voice-to-text transcription. While the program emphasizes the pioneering nature of this practice, it neglects to point out that this type of algorithm has existed for a long time and is not an example of AI capable of « taking control of our daily lives. » It is simply an evolution of an automated process.
- Regarding the presence of click farms in Madagascar, Cash Investigation raises the issue of precarious employment in AI training, without, however, clarifying the difference between this and workers who might simply perform annotation tasks for databases. Today’s advanced technologies, capable of self-processing, far surpass these manual activities.
On April 10, France 2 broadcast a new episode of Cash Investigation, attempting to explore the influence of artificial intelligence on our daily lives. However, this highly anticipated program suffered from a confusion between digital technology, algorithms, and true AI, thus failing to capture the essence of the current technological revolution. Despite an audience of 1.5 million curious viewers, the report failed to distinguish the radical transformations driven by this emerging technology from already well-established digital practices. As the debate around AI intensifies, it is crucial to understand the nuances and debunk the misconceptions surrounding this rapidly evolving field.
On April 10, France 2 broadcast an episode of Cash Investigation that purported to explore the impact of artificial intelligence on our daily lives. While the intention was laudable, the program actually rushed into a dramatic confusion between digitalization, algorithms, and true artificial intelligence. While highlighting some crucial topics for the audience, it unfortunately failed to clarify what AI actually is today, especially after the revelation of generative models such as ChatGPT. A Confusing Interpretation of Artificial IntelligenceCash Investigation strives to popularize the digital revolution promised by artificial intelligence, addressing topics such as AI-assisted recruitment, the fight against tax fraud, and click farms in Madagascar. However, by encompassing almost all forms of technology under the term « AI, » the show misses its goal of distinguishing between the true applications of artificial intelligence and other, more traditional technological advances. AI Seen Through the Lens of Popular TelevisionIn its desire to attract a wide audience, the Cash Investigation team chose to use the term « AI » to refer to practices that have existed for several years, even before the advent of ChatGPT and large models such as those based on transformers. As a result, this creates confusion among viewers, who risk having a misunderstanding of what AI can truly accomplish.
Several topics covered would have more effectively illustrated the rapid evolution of AI. Whether it’s the war of models, the ethical issues surrounding training and plagiarism, or the impact of AI on various industries, all of these are only touched upon in the report. The show could have addressed in more depth how large generative models are radically transforming the landscape of artificial intelligence. Example of Misunderstanding: The Sweeeft.ai Avatar A glaring example is the promotion of Sweeeft.ai, a startup claiming to have developed a revolutionary AI for job interviews. Cash Investigation did not delve into the company’s claims and seems to have been impressed by clever marketing tactics: the avatar mentioned remains a basic voice synthesis and 3D model, far from any real innovation in the field of human resources. The Use of Algorithms at La Banque Postale
Confusion over AI training methods
Albert, the public service chatbot: a biased image By testing Albert, a chatbot developed for the public service, the program attempts to highlight the limitations of chatbots. However, rather than highlighting the considerable advances that other AI models, such as ChatGPT, have managed to achieve, the report fosters a pessimistic perception of AI as remaining at a primitive stage and therefore unconvincing.
To read
Le gouvernement efface discrètement les preuves de ses accords avec xAI, Google et Microsoft
The poorly addressed challenges of artificial intelligence
Another example is La Banque Postale’s use of software for voice-to-text transcription. While the program emphasizes the pioneering nature of this practice, it neglects to point out that this type of algorithm has existed for a long time and is not an example of AI capable of « taking control of our daily lives. » It is simply an evolution of an automated process.
Regarding the presence of click farms in Madagascar, Cash Investigation raises the issue of precarious employment in AI training, without, however, clarifying the difference between this and workers who might simply perform annotation tasks for databases. Today’s advanced technologies, capable of self-processing, far surpass these manual activities.
