DeepSeek &#039, s breakthrough emboldens open-source AI models like Meta&#039, s Llama

  • DeepSeek’s accomplishment represents a win for open-source artificial intelligence concepts such as Meta’s Llama, business specialists told CNBC.
  • Seena Rejal, chief commercial officer of AI company NetMind, told CNBC DeepSeek shows that open-source AI is” no longer just a non professional research initiative but a practical, flexible alternative to shut models” like OpenAI’s GPT.
  • There are downsides to open-source AI, nevertheless— including the risk of computer abuse.
Omer Taha Cetin | Anadolu | Getty Images
The potent new artificial intelligence type developed by DeepSeek is a win for open-source variations of the technology from companies like Metaid, Inc., Databricks, Mistral and Hugging Face, according to industry professionals who spoke with CNBC.

Last month, DeepSeek released R1, an&nbsp, open-source argument design that says to rival the effectiveness of OpenAI’s o1 model&nbsp, using a cheaper, less energy-intensive process.

Due to concerns that the development may result in lower spending on high-performance processing equipment, Nvidia and additional chipmakers ‘ market values dropped.

A Chinese AI laboratory called DeepSeek aims to create significant language versions with the goal of achieving artificial general intelligence, or AGI. It was founded in 2023 by Liang Wenfeng, co-founder of AI-focused numerical hedge fund High-Flyer.

AGI loosely refers to an Artificial that performs duties on par with or above individual intelligence.

What is open-source AI?

Since OpenAI’s ChatGPT first appeared in November 2022, AI scientists have been working diligently to comprehend and advance the improvements of the basic big speech type technology that supports it.

Open-source AI has become a key area of study for some laboratory. Open source refers to application whose source password is made freely available online for adaptation and reproduction.

Numerous businesses, including tech companies like Meta, have been betting on open-source as a means to advance the technology while also sharing essential improvements with the wider research group.

How DeepSeek made empty cause more accessible

DeepSeek’s scientific breakthrough has just made the case for open-source AI types stronger, according to some tech professionals.

Seena Rejal, chief commercial officer of AI company NetMind, told CNBC the Chinese firm’s success shows that open-source AI is” no longer just a non professional research initiative but a practical, flexible alternative to locked models” like OpenAI’s GPT.

” DeepSeek R1 has demonstrated that open-source models can achieve state-of-the-art performance, rivaling proprietary models from OpenAI and others”, Rejal told CNBC. This challenges the notion that innovation in this space can only be dominated by closed-source models.

Rejal isn’t alone. Yann LeCun, Meta’s chief AI scientist, said DeepSeek’s success represents a victory for open-source AI models, not necessarily a win for China over the United States. A well-known open-source AI model called Llama is the brainchild of Metaa.

” To people who see the performance of DeepSeek and think: ‘ China is surpassing the U. S. in AI. ‘ You are reading this wrong. The correct reading is: ‘ Open source models are surpassing proprietary ones ‘”, he said in a post on Linked In.

Read more DeepSeek coverage

” DeepSeek has benefited from open source software and research ( such as PyTorch and Llama from Meta ). They based their new ideas on the work of others and developed them from scratch. Everyone can benefit from their work because it is published and open source. That is the power of open source software and research.

Open-source AI going global

China has turned to open-source technology to increase the appeal of its AI models after being denied by Washington for access to the advanced chips needed to train and run them. In order to promote innovation and spread its use, many Chinese companies are pursuing open source models, including DeepSeek.

However, China is not the only country that is using open-source technologies for AI success. In Europe, an alliance of academics, companies and data centers have partnered on developing a family of high-performing, multilingual large language models, called OpenEuroLLM.

The alliance is led by Jan Hajič, a renowned computational linguist at Charles University, Czechia, and Peter Sarlin, the co-founder of Silo AI, an AI lab that was bought by U. S. chipmaker AMD last year.

In order to reduce reliance on Silicon Valley, nations are encouraging investment in their own domestic AI labs and data centers as part of a wider push for” AI sovereignty.”

What’s the catch?

There are downsides to open-source AI, however. Experts warn that, although open-source tech is a good thing for innovation, it is also more prone to cyber exploitation. That’s because it can be repackaged and modified by anyone.

Sam Altman: OpenAI has been on the 'wrong side of history' post-DeepSeek
Cybersecurity firms have already found flaws in DeepSeek’s AI models. R1 contained critical safety flaws, according to research released by Cisco last week.

Cisco’s AI safety research team claims it received R1 to provide affirmative responses to a number of harmful prompts from the well-known HarmBench” with a 100 % attack success rate.”

‘ DeepSeek R1 was allegedly trained using a fraction of the resources that other frontier model developers use to create their models,” the report claims. However, it comes at a different cost: safety and security”, Cisco researchers Paul Kassianik and Amin Karbasi wrote.

Data leakage is also a concern. Data processed by DeepSeek’s R1 model via its website or app is sent straight to China. Chinese tech companies have long been plagued by allegations that Beijing spies on Western governments and individuals using their systems.

” DeepSeek, like other generative AI platforms, presents a double-edged sword for businesses and individuals alike”, said Matt Cooke, cybersecurity strategist EMEA at Proofpoint. ” While the potential for innovation is undeniable, the risk of data leakage is a serious concern”.

Although DeepSeek is relatively new and will take some time to learn about the technology, Cooke continued,” What we do know is that handing attackers a loaded weapon or sensitive company data into these systems is what we do,” she continued.

NetMind’s Rejal told CNBC that open-source AI models introduce cybersecurity risks which businesses need to consider, including software supply chain attacks, prompt jailbreaking and so-called “data poisoning” events that try to introduce biases or harmful outputs.

WATCH: Why China’s DeepSeek is putting America’s AI lead in jeopardy

Why China's DeepSeek is putting America's AI lead in jeopardy

Leave a Comment