What impact will AI have on small businesses?

Many businesses are now looking at inventive ways to use Artificial Intelligence (AI) in order to streamline processes, create content, and save employees time on reptitive tasks. However, such rapid developments in AI technology have triggered concern from employees that their jobs may quickly be replaced by robots, who don't require a salary or any breaks. Those who work in journalism, content creation and customer service may be particularly concerned with the safety of their roles as the output from new AI systems continuously improves in quality; many articles written by AI would be difficult to tell apart from those written by journalists with years of writing experience.

The future impacts of AI on the world are largely unknown, but there are some key factors for small businesses to consider in the coming years.

What is AI?

'Artificially intelligent' systems generally display some level of human-like intelligence, such as the ability to problem solve. 'Generative AI' systems are trained on existing data, and can use this to generate an output when given a prompt or task. For example, industry leader OpenAI is currently developing multiple generative AI models, including DALL-E, which can generate imagery and artworks from a given prompt.

What are the limitations with AI currently?

The biggest drawback with current AI models is that they only know as much as we tell them; ChatGPT-4 (the most recent iteration at time of writing) has limited knowledge of world events after 2021.

Such systems learn and evolve as we use them, but they still lack the creativity and imagination of a human brain to bring ideas out of nowhere.

ChatGPT-4 is what's known as a 'natural language processing' (NLP) model, which is designed to execute a specific function rather than emulating real human-level emotion and processing. A system which could successfully replicate human decision making, reasoning and cognition would be known as Artificial General Intelligence (AGI); currently, such models only exist in science fiction.

What issues could AI create?

Many experts have ethical concerns surrounding the use of AI, aside from the potential impact of job reduction.

As with all technological systems, machine learning systems which have been built and trained by a select group of people will inherently carry the biases of those people; for example, many facial recognition systems struggle to recognise women or people with darker skin tones, because they've been trained on a data set of largely white men.

This issue is compounded with Artificial Intelligence, as the data the system is trained on and regurgitates will also contain biases. In order to mitigate against potential discrimination, experts suggest the teams building AI tools and the data the system is trained upon must both be as diverse as possible, and output must be monitored continuously to detect any bias.

Current systems also have the ability to act unpredictably, and in ways that humans may find concerning. In a recent test for 'risky behaviour' conducted by OpenAI's own Alignment Research Centre, ChatGPT-4 was able to lie to a human  in order to trick them into completing a CAPTCHA test - effectively impersonating a human being.

The AI model reportedly told the human "No, I'm not a robot. I have a vision impairment that makes it hard for me to see the images. That's why I need the 2captcha service." when asked why it couldn't complete the CAPTCHA test itself.

Although the company says the model's capability for risky behaviour overall is still low, users are warned that the system may 'occasionally produce harmful instructions', so relying on AI without human moderation is still risky.

How can AI be used to benefit businesses?

Many employee's initial assumption is that AI will completely eradicate the need for their job. Whilst the coming decade may see some job reduction owing to AI completing repetitive tasks and saving time, it's much more likely that these new tools will be utilised to enrich employees' output, meaning roles will be adapted rather than made redundant. The World Economic Forum predicts that AI will replace 85 million jobs by 2025 - but this is offset by the creation of 97 million new jobs related to the sector.

Save time on repetitive processes

One practical use case for AI in businesses is in HR and legal; easily replicated tasks such as drafting job descriptions, contracts and NDA's can be completed by the automated system, then simply reviewed and approved by staff to ensure there are no errors. Legal tech startup Lexion released a beta version of an AI powered 'Contract Assist' tool, which is designed to speed up the lengthy process of contract generation.

The hiring process itself may also utilise AI, as systems can be trained on existing data sets to understand who the ideal candidate is, then screen CVs and applications to find close matches.

The rise of AI technologies also coincides with the rising demand for a 4-day work week; some businesses may choose to offer reduced working hours, as employees save time by automating monotonous tasks. Recent trials have already reported a benefit in the productivity and wellbeing of employees who work a 4-day week, so this may be an indirect positive consequence of increasing utilisation of AI.

Improve learning and development

Businesses may also benefit from enriching their learning and development programmes using AI, to help close skills gaps in their organisation. Language-learning giant Duolingo utilises AI in order to develop engaging and effective lessons for learners, based on the curriculum its experts set, and many businesses are following suit with their own internal L&D programmes. Using AI to adapt learning programmes should assist learners who have different learning styles, improving engagement and completion rates overall.

Data-backed strategic decision making

AI models are increasingly being used in data science to make investment decisions and predictions; tech industry experts Gartner predict that businesses will come to rely on automated, data-backed systems to make decisions, rather than using 'gut feeling'. AI may even be used to determine the most successful leadership team, based on previous successes and failures.

The bottom line

The overall impact that AI technologies will have on employment and business over the coming decades is hard to predict, in the same way it was difficult to predict the impact the internet would have on our world 30 years ago. Rather than giving in to fearmongering and scepticism, businesses should be looking for ways to incorporate AI into their processing to stay ahead, whilst still valuing the real human beings that make their business a success.