A guide to what Generative AI is and why it’s important.

Michael-Appleby-headshot
Posted: 24th August 2023 Michael Appleby, Managing Director - Altair Consultancy

Generative AI has developed at an exponential speed over the last few months, and we are all still learning about the potential impacts it will have, both positive and negative.

Because of this, we have launched our own AI focused series. The series, which will consist of multiple articles, will give some focus to AI and what impact it will have on the housing sector as a whole, as well as specific functions in organisations such as governance, HR etc.

We are all still learning about Generative AI, what it is and why it is different to previous versions of AI that we may be familiar with. This article sets out a short guide to act as an introduction to the topic, as well as to be aware of. This also includes some of the potential applications we may start seeing in housing, local authority, and related sectors.

What is Generative AI and why is it different to previous versions of AI?

Traditional forms of AI are based on fixed and pre-designed workflows and rules (for example, think of a chatbot agent on a customer facing website which can only provide pre-determined responses). Whilst useful, this meant that traditional forms of AI could only be applied to a narrow range of specialist activities, and could only work within pre-designed parameters, with no freedom to venture away from that activity.

Generative AI is a term used to describe the latest version of AI, which is based on machine learning algorithms. Generative AI has been trained on huge amounts of data (text, image etc.), it uses that data to make predictions on how it should respond to any prompt or request (the responses can be in many different forms e.g., text, images, music, video etc.). This provides Generative AI with freedom to respond in any way to any request and it is not limited by pre-designed responses or workflows.

Therefore, the key difference between Generative AI and traditional AI, is that the responses that Generative AI provide are unique / bespoke, and it can be applied to virtually any topic.

Why is Generative AI attracting so much attention now?

There are four main reasons for this:

  1. The capabilities of Generative AI have increased far quicker than anyone was predicting 12 months ago.
  2. The user uptake developed rapidly – it only took 2 months for Chat-GPT to gain 100m users worldwide (compared to 55 months for Spotify and 70 months for Uber).
  3. The interest and research in Generative AI are increasing exponentially with estimates now of over 10,000 research papers being published monthly on new AI tools, applications, developments, uses etc.
  4. The tools are easy to use and largely free / cheap to use – none of the most popular versions of Generative AI require any training to use, they can be accessed on any phone / laptop, and most include either a free sign-up or very low monthly subscriptions. This means that the tools are accessible and available to everyone.

It is because of these reasons, that Generative AI interest and the use of Generative AI is now everywhere. Unlike other forms of technology solutions, anyone can access them at minimal cost.

As I referenced in a previous article, this is resulting in a ‘bottom-up disruption’ where pretty much every organisation will already have staff using AI in their day-to-day work, unbeknown to their managers / executives. This is creating lots of new opportunities, but also challenges to manage.

What activities can Generative AI complete?

The simple response is ‘pretty much anything a competent professional can do’.

Now that isn’t strictly true, as Generative AI doesn’t (yet) operate in the physical world, at present it only reacts to prompts, rather than be pro-active, and it doesn’t have a lot of features that we have as humans (nuance, humility, emotions etc.). But as a general overview of what activities AI can complete, the phrase works.

Some general groups of activities and potential applications within the housing sector include:

  • Drafting / content creation – It can help draft letters, reports, or starter for 10 documents. It could draft a communications message to staff for a CEX in a housing organisation, a sales team could use it to develop new sales documentation for a housing scheme, or a housing officer could use it to draft a report analysing data on customer complaints.
  • Complete research – Although some AI tools only have access to data up to certain timescales (e.g., Chat-GPT doesn’t have access to any data after Sept-21), it is able to complete secondary research of data sources available before that time. It could, for example, complete research to understand what the housing need is in Greater London.
  • Language translation – Generative AI can instantly translate information in any other language – in text and sound / voice formats. In a housing context, this has massive potential for positive applications such as engaging with diverse customer groups.
  • Problem solving / diagnosis – A Generative AI customer facing chatbot can help a customer resolve a range of problems or questions e.g., resolving lower-level queries or helping to diagnose repairs problems.
  • Critique and QA work – It can review work and not only check and correct grammatical errors, but also provide hints and tips or feedback on how to improve the structure, content or positioning of a report.
  • Coding – AI now enables us all to become coders, it can convert ideas (that you give it) into coding which can be used to develop or improve software.
  • Summarise – An assistant could use AI to summarise long reports into executive summaries. AI can also be used to record meetings, produce minutes, and recommended actions.
  • Creativity – AI can be used to support in the initial stages of a creativity process. This could include workshopping new ideas, or even using image generation AI to develop early concepts for designs of estates / houses etc.

This isn’t an extensive list, but it starts to give some idea of the potential applications within organisations. It is also important to note that Generative AI is more than ‘just a chatbot’. And whilst it is entertaining and can complete tasks such as talking like a pirate, creating a poem about cheese on a scooter, or creating jokes, it also already has a range of practical and useful applications in our day-to-day lives.

Although the use of Generative AI is already increasing, it’s widespread use won’t be implemented overnight, and some industries will be more affected than others. Organisations such as McKinsey and Gartner predict that the first wave of impact is most likely to be felt in applications such as customer services, marketing, sales, and any other functional process which is high volume.

What are some of the key risks / challenges that I should be aware of with Generative AI?

Due to the sudden uptake in interest, and unregulated / uncontrolled use of Generative AI, the risks and challenges shouldn’t be underestimated. Some of the key areas to be aware of include:

  • Bias – Due to the training data provided, most models produce biases which will impact on the output they provide. Without proper control or additional QA, this could lead to issues in a range of applications, not least in recruitment.
  • Confidentiality / privacy – The tools create new challenges around GDPR, the use of data and Intellectual Property. There is a risk that employees could upload data onto an AI tool, which would effectively mean that the information has been published and, in theory, could then be accessed by other users.
  • Quality – The outputs of AI tools are impressive, but they do need to be checked. The tools have a propensity to ‘hallucinate’ information (i.e., create false data), which without additional checks, could lead to various problems.
  • Assurance in information – AI can instantly generate reports, letters etc. covering a wide range of content areas – to a high standard. It will be easy for individuals to use AI tools to speed up working approaches. This may, however, result in a negative impact on how well individuals can explain the reports that they have developed and give assurance on the information included.
  • Security (including fraud and cybersecurity) – It is likely that individuals will use all forms of Generative AI (including voice-cloning) to complete fraudulent activities and increase the speed / sophistication of cyber-attacks. Housing organisations will need to be particularly tuned into this risk to a) protect vulnerable customers and b) ensure that their organisations IT security is robust enough to withstand attacks.

Again, this isn’t an extensive list of challenges and risks to consider but gives a starter for what should be considered. The challenges also aren’t unusual, and many apply equally to a human completing a similar task.

What should I be doing now about Generative AI?

In this article, I am not advocating that there should be a rush within the housing sector to make significant investments in developing bespoke Generative AI tools.

But, if you are a Board Member or Executive of an organisation within the sector, then there are some things that you should be doing now, including:

  • Increase your own / your Board or Executive Team’s awareness of Generative AI and the potential positive and negative impacts / applications.
  • Understand the potential impact of AI on the sector and on the organisation over the medium and longer term.
  • Raise its profile on the agenda for discussion.
  • Assess the immediate risks / issues that uncontrolled use of AI could create in the organisation.
  • Understand the existing usage / uptake of tools from staff within your organisation.
  • Seek to put a policy position in place to provide some ‘guiderails’ for staff across the organisation, detailing how they can or can’t use AI in their work.
  • If you want to start using Generative AI, identify some initial use cases and pilots to test its impact.

A good starting point would be to consider Generative AI as part of strategy discussions that you are having with the Board and Executive Team, to upskill in understanding, begin the horizon scanning process, and determining the impacts to be aware of.

This is a topic which is only likely to increase in visibility and impact over the coming months, so getting a good understanding of Generative AI now will pay off.

For more information about Generative AI or the impact it will have on the sector / housing organisations, or to include a discussion on AI as part of a Board or Executive strategy session, please contact:

Michael Appleby
Managing Director – Altair Consulting

Latest News

See all news