Article: Opportunity, risk and future trends in generative AI

It’s hard to believe ChatGPT launched just over one year ago. Since then, the platform has gained nearly 200 million users and kicked off a high-speed arms race where Google, Meta, Microsoft and other Silicon Valley players have urgently reorganised around AI.

A 2023 KPMG survey found 65% of executives believe generative AI will have “a high or extremely high impact on their organisation in the next three to five years,” highlighting opportunities around productivity (72%), better ways of working (65%) and innovation (66%).

What will generative AI mean for tech professionals and the wider technology workforce? Ahead of the 8th annual Women in Tech Fest (20-22 February 2024 in Melbourne), we asked a high-powered panel of experts for their opinions on the opportunities, risks, workforce trends and predictions about generative AI.

What are the biggest opportunities of generative AI and how can tech professionals capitalise on these?

Professor Mary-Anne Williams (AAAI (Fellow), Michael J Crouch Chair for Innovation, Deputy Director - UNSW AI Institute, UNSW Business School and Affiliated Faculty at Stanford University) believes we are witnessing a transformative era where generative AI revolutionises content creation, personalisation and productivity, sparking innovation across every industry.

“Generative AI technology is not just a tool; it's a catalyst for new forms of expression and business models,” she says. “To capitalise on these opportunities, tech professionals must focus on skill development in AI and machine learning, identify innovative applications in their respective fields, actively engage in the AI community for collaboration and stay abreast of the latest developments.”

Professor Hind Benbya (Deakin University, Founding Director of the Centre for Artificial Intelligence and the Future of Business and Fellow of the Oxford Internet Institute), notes that despite the technology being in its infancy, its ability to generate new content and perform diverse tasks that were traditionally reserved for humans can provide significant productivity improvements for tech professionals.

“There are myriad benefits currently being harnessed by tech professionals including consultants, software developers, customer engineers and data architects.” Benbya says. “For instance, the benefits of GenAI include its ability to analyse extensive datasets, provide recommendations, generate, or refine content such as code or document summaries. Examples include software engineers using GenAI to complete their coding tasks faster, data analysts and consultants using GenAI to conduct analyses, synthesize diverse results and produce summaries over shorter timeframes.”

For Sarah Carney (National CTO, Microsoft ANZ), it is all about finding the right use cases for generative AI. “The use cases are really only limited by our imaginations. From deeply personalised learning plans through to insight-filled healthcare, there are so many ways in which generative AI can make all our lives fundamentally better.

“When there are so many things you could do, you need to find the right place to start,” Carney says. “Brainstorming or ideating a list of use cases is the first step, but then mapping them out on a grid that helps you identify the use cases that will offer the highest return or level of benefits versus the difficulty of development of implementation. Getting that balance right will help tech professionals get some early runs on the board, as well as building critical skills, which will help get executive and financial support for more challenging use cases.”

What are the biggest risks with generative AI and how can tech professionals mitigate them?

“Generative AI can amplify existing AI risks that organisations are already grappling with,” says Benbya. “Inequity means that the outcomes of the algorithm can put certain groups at a disadvantage. Accountability refers to the inability to define who takes responsibility if the decisions taken by algorithms affect humans in diverse ways. But there are also new risks inherent in GenAI including accuracy and confidentiality.

“GenAI systems such as large language models (LLMs) can write plausible sounding but incorrect answers,” Benbya continues. “So human oversight is necessary not only to ensure that GenAI is actually being trained with the right datasets but also to fine-tune and validate the output generated for accuracy. Another emerging risk is related to confidentiality and data privacy. Users may inadvertently share sensitive information with the GenAI system, which is subsequently stored and potentially reintegrated into a third-party’s system. So, it’s important to use LLMs in a private environment and to create sandboxes and safe spaces where they can analyse their own data without the risk of it being ingested into models and then being made available publicly.”

“Just because we can, doesn’t mean we should,” says Carney. “Think about that as you look at the possible use cases for generative AI. What are the potential unintended consequences? Would you be impacting the quality or access to services or information if you introduced generative AI into your organisation or system? I see most organisations thinking deeply about their approach to Responsible AI, which is probably the single best thing you can do in order to help you identify and mitigate potential risks. Each organisation has a slightly different approach to it, but most have a focus around fairness, security, access, transparency, inclusiveness and reliability of the systems. This helps you consider risk from a number of different angles before moving ahead with implementation.”

“The one thing I see many organisations missing when they deploy a new generative AI product or system is a feedback mechanism,” she adds. “How are you going to know if the system is doing what you expect it to do if the users can’t provide feedback easily when things go wrong? AI doesn’t stand still. It changes over time, so you need to know if things have changed by giving people a way of letting you know.”

Williams notes that while the benefits are substantial, generative AI has risks, including safety, privacy, legal, regulatory and ethical challenges, security vulnerabilities, and quality control issues. “To mitigate these risks, it's crucial to establish and adhere to legal and ethical guidelines, implement robust security measures, consistently ensure quality in AI outputs, and remain vigilant about regulatory compliance, anticipating and adapting to legal changes,” she says. “It's about striking a balance between innovation and responsibility.”

How will generative AI impact the tech workforce and what should tech professionals do to changing workplaces and roles?

According to Williams, the impact of generative AI on the tech workforce is twofold: it's transforming existing jobs and simultaneously creating new opportunities. “This shift necessitates a continuous learning mindset, adaptability to evolving roles, and a strong focus on uniquely human skills,” she says. “As we integrate AI into our workflows, the collaborative potential between human and machine intelligence is immense, opening new avenues for creativity and efficiency that can drive new value generation and innovation.”

“We see the impact happening already through things such as copilots,” Carney says. “It doesn’t matter what role you are in, there is a copilot to help you! For tech professionals, this could be a coding copilot that can help you create your code faster, make suggestions of new code, or even explain what a piece of code does. The biggest challenge will be finding the tasks you can quickly offload to your new AI assistants, versus those where there is value or speed in you completing them yourself. Practicing that delegation and evolving your approach over time will help you adapt to the changing technical landscape.”

Benbya believes that rather than replacing jobs, the potential of GenAI lies in supporting workers by performing certain time-consuming tasks, thereby improving business processes, and enabling the creation of new products and services. “Furthermore, GenAI allows employees with varying degrees of experience and skills to perform comparable tasks,” she adds. “For example, GenAI can assist less-experienced programmers to generate initial code, analysts to conduct different analyses, and tech consultants to provide insightful recommendations to clients.”

This mean GenAI will put some low-skilled jobs at risks of being replaced by AI, but at the same time it will create new jobs.

“Understanding where GenAI can be used as an assistant in certain tasks, its potential, its limits and where it can provide a competitive edge will be important,” Benbya says. “But we need to recognise that these tools are still evolving. Gradually, GenAI will become embedded in most tools so it’s important to continually test and learn how to use it effectively. Leaders need to define clear policies and use cases to ensure transparency and guard against misuse.”

How do you see this technology evolving in the coming years?

The space is evolving so fast that it is hard to even speak to the next few months, let along the next few years. But as the National CTO of Microsoft ANZ, Carney is well-positioned to make the following predictions for generative AI:

  • Commercial grade multi-modal models: “These are models that you can put anything into and get anything out of,” Carney says. “We see these starting to emerge, but the next big evolution will be the ability to put some words and a sound byte into a model and get a full 4K video as an output.
  • Small Language Models: “These are already showing signs of being as precise and capable as the larger models, but with far fewer parameters, which is great news for compute and power needs, but also for organisations looking to build their own very precisely trained model.”
  • Combinatorial use cases: “We will see generative AI being built into systems all around us. Whether that is Mercedes testing it in millions of vehicles to give you a more conversational experience that taps into your application ecosystem, or extended reality systems that you can chat with to get updates, assistance and information, all in your virtual spaces.”
Williams anticipates the evolution of generative AI will be both technological and cultural as we adapt to and integrate these tools into our societal fabric. “The future of generative AI is incredibly promising, with research leaning towards improved AI governance, advanced language understanding, and ground-breaking applications in healthcare, biotechnology, transport and education,” she says. “As this technology evolves, it will likely become more integrated into everyday technologies and accessible to a wider audience.”

What tips would you give to tech juniors trying to harness generative AI effectively?

“Just play with it!” advises Carney. “Generative AI is still so new that there is a wonderful opportunity to build deep knowledge in the area very quickly and be some of the first to really shape the future of the technology. For example, prompt engineering was a phrase that really only entered our consciousness less than 12 months ago. What an amazing opportunity to learn and develop skills in this field while it is still nascent.”

“The technology is evolving so quickly that users need to really bring their agile mindsets,” continues Carney. “The way we build and deploy generative AI systems today has completely changed from just six to eight months ago. You have to be OK with throwing away what you worked on, or just keeping parts of it and starting again. What you have gained in building something is knowledge that will help you be even more successful with the next project.”

Benbya adds that these tools require significant human oversight and new skills to be used effectively. “Tech professionals need to learn how to work alongside these new tools in a kind of a ‘new way of human-machine collaboration’ while being clear on the limitations of GenAI. Learning gradually how to develop instructions to generate the best output from prompts is an important skill, as well as sense-checking results for accuracy.”

Finally, Williams’ advice for those at the outset of their tech careers is to build a strong foundation in AI basics, engage with AI tools hands-on, keep up with industry trends, understand the legal, regulatory and ethical implications, cultivate a robust professional network and connect with the AI innovation ecosystem. “A blend of technical knowledge, practical experience, and ethical consideration will be key to thriving in this rapidly evolving field.”

If you would like to learn more about how generative AI will affect tech professionals and the wider technology workforce, please join us at the Women in Tech Fest 2024 on 20 – 22 February where you will also be able to hear more from Sarah CarneyMary-Anne Williams and Hind Benbya as well as over 30 other inspirational tech leaders. Learn more.

To access the detailed conference program, download the brochure here.