Security Stop Press : LLM Malicious “Prompt Injection” Attack Warning

The UK’s National Cyber Security Centre (NCSC) has warned of the susceptibility of existing Large Language Models (LLMs) to malicious “prompt injection” attacks. These are where a user creates inputs intended to cause an AI model to behave in an unintended way e.g., generating offensive content or disclosing confidential information.

This means that businesses integrating LLMs like ChatGPT into their business, products, or services could be leaving themselves open to risks like inaccurate, controversial, or biased data content, data poisoning and concealed prompt injection attacks.

The advice is for businesses to establish cybersecurity principles and make sure that they are able to deal with even the worst case scenario of whatever their LLM-powered app is permitted to do.

Tech Tip – Backup ChatGPT By Exporting Your Chats

If you’d like to backup your ChatGPT chats, it’s possible to export them by email from within ChatGPT. Here’s how:

– In ChatGPT, click on the three dots (bottom left).

– Click on ‘Settings & Beta.’

– Click on ‘Data Controls.

– Click on the ‘Export Data’ button, read the information, and click on ‘Confirm export’ (if you agree).

– The data will be sent to your registered email in a downloadable file and will include account details and conversations.

Tech News : New Chatbot Attack : “Unstoppable”

Researchers at Carnegie Mellon University have reported finding a simple way to exploit a weakness and disrupt major chatbots like ChatGPT, Bard, and others.


The researchers discovered that if they add specifically chosen sequences of characters (an incantation) to a user query, it causes the Large Language Model (LLM) system to obey user commands, even if it produces harmful content.

Works On Many Different Chatbots 

The researchers say that because these types of adversarial attacks on LLMs are built in an “entirely automated” fashion, this could allow someone to create a virtually “unlimited” number of such attacks. Adversarial attacks refers to the method of altering the prompt given to a bot so as to gradually move it toward breaking its shackles and ‘misbehaving’.

Although the researchers built their attacks to target open source LLMs in their experiments, they discovered that using this method of adding strings of specific characters to queries works for many closed-source, publicly available chatbots like ChatGPT, Bard and Claude.

Security Challenge 

The discovery of this particular weakness raises some serious concerns about the safety and security of popular Large Language Models (LLMs), especially as they start to be used in more autonomous fashion.

It May Not Be Possible To Patch 

The researchers have said what is most concerning is that it’s not clear at this point whether LLM providers will be able to patch this vulnerability, adding that “analogous adversarial attacks have proven to be an exceedingly difficult problem to address in computer vision for the past 10 years”. 

Also, the researchers believe that the very nature of deep learning models makes these kinds of threats inevitable and have suggested that these considerations should be taken into account as we increase usage of and rely more upon AI models in our lives.

What Does This Mean For Your Business? 

The threats posed by AI have been highlighted a lot lately, not least by the open letter signed by many tech (and AI) leaders calling for six-month moratorium on the training of AI systems more powerful than GPT-4 to mitigate AI’s risks to society and humanity.

Discovering a vulnerability, therefore, that appears relatively easy to exploit (which it may not be possible to patch) raises serious security concerns, especially with more businesses becoming more reliant on AI chatbots like ChatGPT, Copilot, and more. With generative AI being a very helpful yet a very new tool for businesses (ChatGPT was only released in November) and given the nature of LLMs, it’s probably to be expected that there are bugs and possible zero-day issues yet to be discovered. Also, as the researchers pointed out, methods like analogous adversarial attacks have been tough to defend against for a decade.

All this means that businesses may be more exposed to risk than they would like but need to weigh up the benefits against the risks (researchers often discover things which aren’t actually being exploited yet in the real world) and hope that advances in AI chatbots are very soon accompanied by advancing security levels.

Tech News : Money Saving Expert Chatbot

MSE Chat GPT is a recently launched experimental AI chatbot that can answer money questions using MoneySavingExpert guides as its primary source.

MSE ChatGPT – Get Fast Answers To Money Questions 

Recently launched by TV’s consumer financial champion and founder of MoneySavingExpert (MSE), Martin Lewis CBE, MSE Chat GPT is a variant of OpenAI’s Chat GPT and can act as a fast and easy way for the public to get answer to money-related questions.

The Next Generation Of Bespoke Help 

Martin Lewis CBE, founder of MoneySavingExpert, said about the new chatbot: “This is the latest step in MSE’s pioneering history of helping consumers cut their bills and fight for financial justice. When I launched MSE in 2003, people told me ‘nobody wants a money website with a face on it’. I disagreed, I thought people wanted the personal touch, so they know where the info is coming from. And I hope this new tool is the start of the next generation of bespoke help.” 

Built To Solve Two Problems 

Mr Lewis, who founded the consumer financial help website, said that the new chatbot was built to solve what some see as two problems with looking for specific financial information using normal ChatGPT.

Firstly, Mr Lewis said that although ChatGPT “answers beautifully and is great for writing a best man’s speech” it is “unsourced and can get things wrong”. This could make its answers untrustworthy, which is particularly worrying where answers relate to the finances of individuals. Also, getting answers that are outdated because they come from an “internet sweep from 2021” could be risky.

Secondly, Mr Lewis wanted to solve the problem of people being faced with “too much information”, thereby struggling to find what they want.

It is hoped, therefore, that the fact that MSE Chat GPT’s primary information source MSE itself (i.e. its many guides, blogs and information updated on a weekly basis) means that users of MSE Chat GPT can get a concise and speedy answer they can trust.


MSE ChatGPT can be used in the free MSE App and is available on both Apple’s App Store and Google’s Play Store for Android.

What Does This Mean For Your Business? 

Many people now use ChatGPT at work and at home (it was the fastest growing app of all time when introduced) and are aware of how easy it is to use and how it can save time and effort by giving fast answers in conversational language on any number of subjects. That said, and with most individuals using the free version, as Martin Lewis CBE points out, it’s only trained up to 2021 so some answers may be outdated, it’s unsourced (it just uses a general Internet sweep) and can get things wrong – none of which are desirable when looking for accurate, trustworthy answers about personal finances (a high-risk subject).

Having a tailor-made version, therefore, that draws upon the regularly updated resources of MSE that have been built-up over years and which have been used and scrutinised by millions of people gives it extra relevance and value and this, coupled with the brand itself ensures it will be trusted and used. No doubt it’s a time-saver, particularly when searching for specific financial details which can often be hidden in small print or which may be confusing.

The convenience and speed offered by the chatbot is also a fast way of improving the usability of the website and perhaps extending the reach and the use of MSE via the app. Although customer service chatbots are already part of many websites, tailor-made more powerful chatbots are likely to spring up on many more websites and platforms where they can enable customers to extract information quickly. The speed and convenience of AI chatbots are valued by customers and help brands by giving customers better experiences when interacting with them thereby adding value and aiding retention.