You can hardly have missed over the past few days and weeks that there’s a new AI on the block – DeepSeek. Hailing from a lab in China, it has been boasting that it’s just as good as ChatGPT et al but was built at a fraction of the cost. The traders duly hammered the stock prices of various established AI vendors but things got murkier as time went on and details about it were revealed…
You see, it appears that it censors anything to do with the Chinese government and anything else that may result in less-than-favourable views of the Chinese nation. Given the marketing that it’s for everyone to use and that you should use it for your business, how can you trust what it tells you? Not only that but China hardly has the greatest reputation for the privacy of its own citizens so who is going to argue with me that they’ll hold your – and your business – data privacy in any higher regard?
Don’t get me wrong, I hold exactly the same views for the other AI platforms (Apple Intelligence, ChatGPT, Copilot, Gemini etc…) to a certain extent – the potential for harvesting your data for later sale or fake / misleading news isn’t something to be discounted.
There have been claims of theft of intellectual property in the building of DeepSeek which have yet to be proven but perhaps the reputational damage is done? (DeepSeek would obviously have been cheaper to develop if they did indeed copy someone else’s hard work.) Of course, companies fire up their photocopiers the moment that a competitor comes up with a compelling product or service but outright theft isn’t the way to do thing. Build upon things yes, but don’t just copy.
The security of the DeepSeek platform itself is of great concern too. Recent reports tell stories of how critical vulnerabilities were found in its databases and some were even publicly accessible from the internet. Not only that, the iOS App for DeepSeek was revealed to send your queries unencrypted and in plain text! This means that any queries sent to DeepSeek could be intercepted and read with very little effort by someone willing to take the trouble to do so. Yes, Apple has a privacy framework built into iOS that it encourages its developers to use but it does’t enforce their use.
That’s enough of bashing DeepSeek but there’s more to the AI conversation…
As always when talking about AI and its uses the doomsayers start lamenting that jobs will be lost, we’re all going to lose skills and that Skynet will soon be rearing it’s metaphorical head and we’ll all be overrun by Arnold Schwarzenegger lookalikes. But I don’t think that will happen.
If you remember that these are tools for us to use as guides and perhaps inspiration for those lengthy blog posts (not this one I might add!) then your view of them can change. I myself have thrown the odd query at ChatGPT and was quite impressed with the answer it gave, but more importantly it also offered a view to the question that I hadn’t even considered.
Whilst I will be staying well away from DeepSeek and its various troubling ‘qualities’, I will continue to explore the potential uses – and pitfalls – of these new AI overlords tools.
This article was written by a human (probably – JS).
