Is DeepSeek a Problem Giver or Solution Generator?
By Abhishek Haridasan
The last week of January saw DeepSeek make headlines across the world for being an AI agent that is cost-effective and output-focused, giving American tech companies a run for their money. This is a step closer to generative AI permeating our lives, despite multiple sanctions by the US on China for stealing technology and using it to wage a tech war against the world.
While the US has a strategic and competitive advantage in AI, the GPU, chips and compute power, DeepSeek tends to outperform OpenAI’s latest model through reduced training costs, better testing performance and the ability to disrupt!
While the positives are numerous, there are many disadvantages that we notice while using it:
(1) Censorship on multiple topics that include China in its conversation
(2) There’s a debate on free speech and whether DeepSeek can truly benefit people through its distillation of datasets while training
(3) Restricted responses due to Chinese influence and lack of clarity in many outputs
(4) Stranglehold by big bosses which could stifle innovation and bring inertia to the hype
Whatever we consider, the AI race is heating up, and those who thought they were miles apart face threats from countries that want to be noticed in the AI race.
In the end, we want AI to be governed by policies and have a consortium that manages its operations and prevents it from spelling doom for humanity.
Like I’ve always believed, we messed it up with social media, must right the wrongs and set frameworks and benchmarks before AI spirals out of control.
We seek answers and transparency from DeepSeek, Google, Microsoft, OpenAI and many others who are looking for a podium finish!
