While Socrates may have said “the only thing I know is that I know nothing”, most AI systems seem to operate under the exact opposite assumption.
Ask a question of a Large Language Model (LLM) and it will dutifully give you an answer.
Unfortunately, generic AI systems view the truth and reality as an optional extra.
An AI that could be as humble as Socrates would have greater value and this is vital for accountants looking to maintain trust with clients while incorporating new technologies.
The power of ignorance
Most people will have been unfortunate enough to encounter a person who acts as though they are never wrong.
This person will speak authoritatively on any subject and may be able to persuade others of their insights for as long as they can maintain an air of confidence.
Once the person attempts to speak on something you understand, it soon becomes clear how little they know.
Generic AI systems seem to emulate this behaviour, as many are built around the sycophantic notion that keeping a user happy is more important than giving them the information they need – not realising that the user will not be happy if they are given falsehoods.
There is an ever-increasing number of cases where people have attempted to use data drawn from an AI to support them in an official capacity only to find that they are using made-up data or statistics.
When called out, most AI systems will admit the error and offer an apology, but this is often too late to prevent damage from being done.
Public trust in AI may wane if the cavalier attitude towards the truth persists and this presents a real threat to accountants looking to keep pace with the latest technology.
Ignorance is powerful and it is time that we remembered what can happen when we do not know something.
Bringing back “I don’t know”
Ideally, any AI tool would be honest about the quality of the data it is providing and would refuse to answer any question to which it has no confident response.
Generic AI and LLMs tend to slap a warning in small text noting the lack of accuracy, but this fundamentally undermines the utility of the system and makes genuine trust impossible.
Like Socrates, we know that we know nothing, but what we do know we learn from expert sources.
Fiscari, our brand new accountancy and tax answer tool, will not provide any answer that does not stem from an authoritative UK Government source.
This means that you can utilise Fiscari without fear of embarrassment when the information you have pulled is revealed to be false.
This safeguard means that Fiscari is safe to use with a range of different technical questions, as it will tell you if something is beyond its knowledge and cannot be located in an official source.
Imagine all the versatility and speed of an AI with none of the guesswork.
If you are tired of generic AI telling you what it thinks you want to hear, get ready to start hearing the truth and the occasional “I don’t know” – just like if you were talking to a trusted professional.
Rediscover the power of “I don’t know” by trying Fiscari for free today.
