Hype or helpful? Making sense of AI in AT

There’s no doubt about it, Artificial Intelligence (AI) is becoming part of everyday life. Whether you’ve got an AI bestie who answers all your questions, an AI-assisted wheelchair that sees the room around you before you navigate it, or a handy vacuum that uses AI to map your floors and clean them, the uses for AI are coming thick and fast.

With AI-powered Assistive Technology (AT) creating new possibilities for connection, accessibility, and support for people with disability, it’s an exciting time. But how do we know what’s useful and what’s just the latest craze?

When it comes to AI-enabled AT, we need to make sure the products we use are safe, do what we need them to do, and bring value to our lives – and to make decision-making easier, the National Disability Insurance Agency (NDIA) has a framework that can help. It’s called the 'Framework for AI-enabled assistive technologies as supports under the National Disability Insurance Scheme (NDIS)'.

How the framework was developed

The framework was developed by the NDIA in 2022, as part of a collaboration between the Agency and the Australian e-Health Research Centre, the Commonwealth Scientific and Industrial Research Organisation’s (CSIRO) national digital health program. 
The team behind the framework spoke with people with disability and their support networks, along with sector, industry, and government stakeholders, to understand how AI fits into everyday life, and where it can be useful.

Many of those participating in round table discussions reported AI-enabled technology was already helping them to live more independently, stay organised, communicate with others, and maintain their daily routines – and some said it had a significant and positive impact on their quality of life too.

However, there were calls for better information about available technologies, more transparency from manufacturers and suppliers, and trusted, independent advice to help people compare products and make informed choices.

Everyone involved agreed there was a need for a simple and consistent way to assess AI-enabled AT, particularly with new technologies and products rapidly entering the disability sector and promising big things. 

The framework in action

Rather than focusing on the technology itself, the framework puts people with disability at the centre of decision-making, recognising the best product is the one that best fits a person's goals, preferences, environment, and support needs.

The framework is built around six key areas that can help you decide whether an AI-enabled AT support is a good fit.

1. User experience

Is it easy to use?

Technology should work for the person using it. Things to consider include:

• Is it easy to learn?
• Does it help you achieve your goals?
• Is it accessible and suited to your abilities and preferences?

A product that’s difficult to use is less likely to be helpful in the long term. 

2. Privacy and security

Does it protect personal information?

Many AI-powered products collect and store data. Ask questions such as:

• Is personal information protected?
• How is data stored?
• Who can access it?
• Have you given informed consent?

Privacy and security should always be considered before adopting new technology. 

3. Quality

Is it high quality?

Not all products are created equal. The framework recommends asking:

• Who developed the product?
• Has it been tested and validated?
• Is the information used by AI accurate and representative?

Reliable products should be backed by evidence and ongoing quality assurance. 

4. Safety

Is it safe?

Safety goes beyond physical safety. It can also include:

• Emotional wellbeing
• Cognitive load and fatigue
• Social impacts
• Reliability over time

The framework encourages people to consider both the potential benefits and the possible risks of the products they choose.

5. Relative value

Does it provide value for money?

A more expensive product is not always the best option. The framework recommends considering:

• The overall cost of the product
• Training and maintenance requirements
• Expected outcomes
• Whether another solution could achieve similar results

The focus should be on outcomes, not just price. 

6. Human rights

Does it respect human rights?

AI should support inclusion, choice and control. The framework highlights the importance of:

• Fairness
• Transparency
• Ethical use of AI
• Avoiding discrimination
• Compliance with Australian laws and regulations

It’s important to understand when AI is involved in an AT support that you’re considering using, and it helps with making sure your rights are protected. 

Making sure AI-enabled AT is right for you

AI-enabled AT has enormous potential. It could help more people communicate, navigate their community, manage daily tasks, and participate more fully in work, study, and social life.

At the same time, choosing the right technology can be complex. If you have funding for AT and you’re in the market for a new AT support, the framework provides a practical way to think about whether an AI-enabled support is the right fit for your unique circumstances, goals, and preferences.

Most importantly, it reminds us that AT should never be about the technology alone. It should be about helping a person to live the life they want to live.