Is AI OK?

Artificial intelligence, or AI, is a technology that enables computers to simulate human intelligence and problem-solving capabilities. It’s the technology of the moment and is being rolled out everywhere – from automated chatbots on websites to enhance efficiency in customer service, to online apps such as ChatGPT that can answer written questions and write documents (but not this one), and apps that can edit images smartly to apply the head of one individual to the body of another.

But AI has application in many other fields as the December 2024 issue of National Geographic noted. It reported, among other things, how AI is helping to decode the hidden lives of animals. In particular, it’s been set to look at what whales are “saying”, interpreting livestock behaviours and how birds in a tropical rainforest communicate.

As for the workplace, AI can do so much more (Table 1). In terms of security, it can help with data analytics and surveillance; it can offer speech-to-text and text-to-speech converters; it can help target areas of a given market; it can aid data-driven business forecasting; it’s used for fraud detection and risk mitigation; and when it comes to hiring staff it can give an impartial – subject to programming – data-driven employee performance assessment.

In the veterinary and animal world AI has applications too (Table 2). It can help with faster test results, health monitoring and data analysis. AI algorithms can, for example, be fed medical data to learn how to interpret digital images to search for anything out of the ordinary while offering better diagnoses and suggestions for treatment.

In the right hands AI can be a force for good. But in the wrong hands, or used incorrectly, AI can do serious damage to an individual or a business precisely because it cannot apply the nuances of life that we, as humans, understand.

Used incorrectly, AI can do serious damage to an individual or a business precisely because it cannot apply the nuances of life that we, as humans, understand

How AI can benefit the workplace

Streamline processes

None of us are experts in everything so it makes sense to use AI to tap into knowledge and skills that we don’t possess. Consider the spreadsheet. A formula is needed for a given calculation, but we don’t know how to write it. Rather than spending hours tapping away and hoping for the best, AI can, if asked the right question, generate what we need in seconds.

Enhance productivity

Humans have finite concentration spans. Technology, however, doesn’t get tired or distracted. And just as we tire, so AI stays alert and can stay creative. As a result, AI can help organisations to think outside the box.

Think about writing a job advert. A manager can scope out a job description and an advert in an afternoon. But how successfully could that same manager then tailor the advert to different audiences and distinct demographics – or even languages? Again, the correct command to ChatGPT could have the system rewrite the text differently. This could save a small business much time and money in not having to use a third-party recruitment provider.

In a veterinary context, AI can help with record keeping along with the automation of admin tasks such as appointment management, stock management and even billing

In a veterinary context, AI can help with record keeping along with the automation of admin tasks such as appointment management, stock management and even billing. Even better, it can help with dictation and speech recognition, a problem that is particularly troublesome in medical environments.

Improve quality

Just as technology and automation has done much to remove the drudgery from repetitive manual tasks, so AI can help employees focus on the right and productive things that both interest them and benefit the business. Not only will this mean greater levels of efficiency and productivity, but it can also help employee engagement and so lower the odds of staff leaving for better and more interesting jobs; anything that keeps staff on side has to be a boon for practices where recruitment is tough.

In other areas, AI can enhance the client experience of a practice. Chatbots that are common elsewhere can be deployed in a practice environment so that, with the right programming, they can answer common questions and give basic pet care advice if help is needed out of hours.

AI can also help a practice improve its marketing and, by analysing client and pet data, can create personalised marketing profiles.

Watch the risks

AI is helpful but it is far from perfect, and it helps to understand its limitations. Simply put, machines aren’t intelligent and cannot learn. However, they’re great at taking information and interpreting it.

Confidence

By extension, it’s necessary to recognise that using a public AI system means a loss of confidentiality

By extension, it’s necessary to recognise that using a public AI system means a loss of confidentiality. So, uploading something for Chat GPT to look over – a proposal to merge practices, papers, contracts, and so on – means that it’s going to amalgamate information uploaded with what it has. That puts the upload in the public domain where it cannot be secured. Practices need to recognise this and create a policy over its use so as to not breach confidentiality obligations.

The GDPR has relevance here and certain types of personal information will be doubly protected as a breach could be very serious with potential for serious harm to data subjects.

Garbage in, garbage out

As before, AI is good but lacks the nuance that humans can apply – even if the output is grammatically perfect.

Not only can the likes of ChatGPT create something that reads poorly, there’s also the risk that it can inadvertently reuse the content of others and so breach their intellectual property rights.

In the real world, being caught out using someone else’s material, and poorly too, is a surefire way to lose credibility with customers. The only solution is to label AI-created material as such.

Similarly, it needs to be remembered that AI “learns” according to what it’s been taught. That means, for example, if it’s been taught to look at CVs – and the base material relates only to men – then the AI system will more than likely skew output towards men and so discriminate against women. And that, in an employment tribunal, can get very expensive if proven.

And let’s not forget the “live chat” that chatbots offer. Being AI-based until the matter is handed over to a human, they can cause trouble: in one instance a chatbot at DPD was taught by a customer to swear, call itself useless and write a poem about how bad the company was. In many situations, there’s no substitute for human interaction.

AI for decision making

It’s imperative that practices are careful not to rely solely on automated decisions and ensure that AI-generated results accurately reflect the data used

It’s no coincidence that with greater levels of digitisation some rely on AI to search records and provide decisions such as creditworthiness or identity. As a result, it’s imperative that practices are careful not to rely solely on automated decisions and ensure that AI-generated results accurately reflect the data used.

The same applies to AI in staff management. Firms need to use AI carefully with the system playing no part in disciplinary matters or on deciding on matters such as flexible working or any reasonable adjustments. Systems may be used but humans must always decide.

Train staff

With all that’s been noted above, it is key to ensure that those using AI in any form are trained on correct protocol and data protection, confidentiality and content usage; they need to understand that AI is not infallible. In other words, they need to be taught to question when results don’t seem right.

Summary

AI is just another tool that, like others, needs proper supervision. While it can offer much in the way of alternative thinking and efficiency, it also has the power to be harmful. Common sense needs to be applied to its use.