Artificial intelligence is advancing at a giant pace, with new startups and models emerging constantly. One of the most recent to draw attention is DeepSeek, a Chinese startup that has launched two AI models that promise to compete with the giants in the sector. However, its arrival has generated strong controversy, mainly due to misuse. Protect your company with secure AI and avoid risks associated with technologies of dubious origin.

DeepSeek recently announced the launch of two new models: R1, designed for image processing and computer vision, and V3, an update of its base model with improvements in reasoning and multimodal capabilities. With these innovations, DeepSeek seeks to position itself as a serious competitor in the AI market, currently dominated by OpenAI, Google DeepMind, and Anthropic.

Dubious Origin

Unfortunately, experts have identified a series of problems that call into question the legitimacy of DeepSeek's advances

Suspicious similarities with GPT-4 and Claude 3

  • Some researchers have noticed that responses generated by DeepSeek V3 are too similar to those of GPT-4, raising suspicions about possible unauthorized use of OpenAI and Anthropic technology.

    • Surprisingly rapid progress

  • DeepSeek has achieved notable advances in very little time, which has raised doubts about how they managed to develop such a sophisticated model without access to the same resources and datasets as their competitors.

    • Lack of transparency in model architecture

    blog.posts.1.content.subSections.0.lists.2.description

  • Unlike other AI companies, DeepSeek hasn't shared clear details about their training process and the data used, which has increased distrust.

    • Center of Controversy

    The controversy is not limited to lack of transparency. There are three major concerns surrounding DeepSeek:

    1. Misuse of technology

  • DeepSeek has been accused of using OpenAI and Microsoft technology without authorization. It's suspected they used a technique called "distillation" to improve their model by learning from more advanced systems, which could violate OpenAI's terms of service and generate legal problems.

    • 2. Censorship and bias in results

  • Users have reported that DeepSeek avoids answering or censors politically sensitive topics in China, such as the Tiananmen Square protests or the Uyghur situation. This has raised concerns about the model's impartiality and the potential use of AI to reinforce government narratives.

    • 3. Data security and privacy

  • DeepSeek stores user data on servers located in China, which raises concerns about government access to this information. In fact, Italy's data protection authority has already blocked the application and initiated an investigation into its handling of personal data.

    • Impact on the technology market

    DeepSeek's launch has also had a significant economic impact. Uncertainty about Chinese competition in AI has caused approximately $1 trillion drop in the market value of U.S. technology companies linked to AI. This has generated a debate about global competition in the sector and the national security implications associated with the development of AI models in different countries.

    Given this landscape, it's crucial for companies to choose safe, ethical, and reliable AI solutions. And this is where Enric AI makes the difference.

    Enric AI: The safe and ethical AI solution

    Enric AI positions itself as the safest option for companies looking to implement AI without compromising privacy, security, or transparency. Our training and data generation architecture ensures regulatory compliance, avoiding legal and ethical risks.

    1. Models built on verified foundations

    Unlike DeepSeek and other competitors, Enric AI uses Falcon LLMs, open-source models trained exclusively with verified public data from the RefinedWeb Dataset. This means:

  • Zero copyright risk: We don't use private or copyrighted data.
  • Regulatory compliance: Our training process aligns with international standards for AI privacy and ethics.

    • 2. Mandala Architecture: Learning without legal risks

    Our Mandala Architecture technology allows Enric AI models to continuously improve without infringing intellectual property rights.

  • We generate safe synthetic data: This allows training models without relying on third-party content.
  • We avoid bias and ensure transparency: We control AI learning to prevent manipulation or censorship.

    • 3. Customized AI for each company

    Enric AI offers 100% customized AI for each client. This means:

  • Total data control: Each company has its own model, without sharing information with third parties.
  • Guaranteed compliance: We ensure each model is free of protected material or unauthorized private data.

    • Protect your company with secure AI

    DeepSeek has shown that AI without clear regulations can generate ethical, legal, and security risks. Accusations of technological appropriation, censorship, and privacy violations underscore the importance of choosing reliable providers.

    Enric AI offers a safe, transparent, and ethical alternative, allowing companies to harness the full potential of AI without exposure to risks. If you're looking for AI that guarantees security and regulatory compliance, Enric AI is the solution.

    Let's talk! Discover how Enric AI can enhance your business without compromising security or ethics.

    Book a Demo