Insight: Ensuring that the future of artificial intelligence is a positive one
Infobip’s head of product Krešo Žmak was interviewed for Medium’s Authority Magazine to provide his take on the future of artificial intelligence. We dig into the highlights and what needs to happen for AI to fulfill its potential as a powerful force for good.
If you have been to the cinema in the past few months, you will know that there is a new villain in town. Foreign spies and aliens are so last decade, the new threat according to Hollywood is…artificial intelligence!
A bit more on the fantastical end of the scale, but we must also mention the chilling M3gan about an animatronic doll that achieves sentience and goes on a killing spree, a bit like a cuter version of Arnie’s Terminator from the 1984 film. Forty years ago? Now we feel old.
But why has the tide turned against AI? Wasn’t it supposed to be our great savior, doing all the repetitive work and helping us to save time so that we could do more fun stuff, like going to the cinema?
A media backlash
We can attribute the U-turn to a number of factors. Firstly, we had the ChatGPT launch that thrust generative AI into the public domain last year, followed by a number of similar solutions that were perhaps rushed out a bit prematurely. More recently there have been sensational stories about industry insiders breaking rank and criticizing the unregulated development of AI and the dangers it could pose.
The AI expert Geoffrey Hinton recently talked to media about how he left Google due to his fears about the dangers of advancing AI saying, “I don’t know any examples of more intelligent things being controlled by less intelligent things”.
The debate has culminated recently with Elon Musk, Mark Zuckerberg and others having a closed-door meeting with US senators where they described the “civilizational risk” that unregulated artificial intelligence poses to society.
But how much of this is actually true and how much is just hype?
Against this backdrop, our head of product Krešo Žmak was interviewed for Medium’s Authority Magazine (check out the full article here) to provide a calm and rational take on the future of AI.
In this blog we look at the highlights, and what businesses and regulators can learn about using AI effectively and safely.
Infobip’s head of product Krešo Žmak was interviewed by David Leichner for Medium’s Authority Magazine.
The role of AI in product development
Krešo makes the point that AI should not be viewed as a final solution but as a tool that can be used by product managers, data scientists, and developers to create better solutions. By taking a two-pronged approach, businesses can incorporate AI into their own processes to make teams more efficient by automating tasks like documentation creation and backlog grooming and include it in the products that they create.
He goes on to stress that the basis of any successful AI project is collaboration between engineers, data scientists, and CX professionals. Only by working together can they leverage AI to create effective and efficient solutions that acknowledge the importance of people in the whole story by incorporating their creativity and ‘humanness’ that no algorithm can replicate.
Approaching an AI project in this way ensures that a successful outcome means tangible positive benefits for people, without them feeling locked out of the process and unable to scrutinize, evaluate, and ultimately control the role that artificial intelligence plays.
What is AI good at?
- Speeding up successful outcomes: The recent developments in large language models (LLMs), for example the Generative Pre-trained Transformer model that makes ChatGPT so impressively fast, have been a huge step forward in speeding up a myriad of tasks. Until their development, even high-powered computers would need some time to accurately segment vast amounts of data, or recognize intent fast enough to replicate the pace and flow of a conversation between two humans.
- Democratizing automation: The ability to automate highly repetitive tasks is a huge step. Of course, automation has been around for a very long time – the term was first coined in 1946 to describe the use of automatic devices in production lines. But modern AI extends the benefits far beyond industrial applications – anyone with access to a computer or even a smart phone can build automation into workflows and processes.
- Medical applications: Machine learning has become instrumental in the faster diagnosis of diseases, speeding up the development of new drugs, personalizing treatment for chronic patients, and has even helped scientists become more efficient at gene editing, which has exciting potential for the prevention of diseases that can still be fatal.
- Stimulating productivity: Ironically, considering the criticism that has been leveled at AI tools for stealing jobs in the creative industry, LLMs have actually stimulated creativity and productivity in many areas by handling many of the routine daily tasks that took up so much time, freeing up humans to work on creative new solutions for many of the critical challenges that society faces from climate change to power generation and the medical breakthroughs that we have already mentioned.
- A catalyst for further investment: The hype surrounding LLMs has stimulated substantial investments, creating even more opportunities for applying AI in a positive and useful way for the benefit of humans.
What is concerning about the inappropriate use of AI?
- Lack of transparency: If like us you have experimented with ChatGPT and other similar tools then depending on the subject you are dealing with the output can range from biased to completely wrong. By scraping the entire internet for information, invalid or even parody content can be presented as truth. One of the key shortcomings of LLMs is that there is a lack of transparency in data sources. Some solutions that are based on LLMs are working to reduce this risk, for example Microsoft’s search engine Bing displays some of the sources that it has used to generate a response so that the user can decide for themselves whether they are credible. However, it is clear that AI can’t be relied on to police itself for subtly biased or discriminatory content.
- Treating AI as a final solution: As we have discussed already, we shouldn’t always be looking to AI to provide a complete ‘silver bullet’ solution to business and societal challenges. Trying to use AI exclusively in situations that aren’t suitable will lead to failure. Even with the extensive development and investment that has gone into autopilot solutions in the aviation industry, they are not seen as a replacement for pilots, but as an aide. In the same way, those that believe that AI can replace the roles of developers, product managers and even artists, are misguided as AI will never be able to replace the expertise, ingenuity, and adaptability of humans.
- Negative impact on society: The potential for the misuse of AI throughout society is a huge worry. From students getting AI to write their essays, to businesses replacing human jobs, to criminals gaming AI to commit fraud and even influence governments, the threat is genuine. We are not being melodramatic when we say that the inappropriate use of AI will have an impact on the long-term success of the human species.
Is AI really a danger to humanity?
Rather than Hollywood’s fixation with malevolent AI that tries to rise up and annihilate the human race, the most immediate and concerning threat is the potential for human replacement.
In the Medium article, Krešo makes the very valid point that by relying on AI to do a growing amount of our work will ultimately hinder innovation and cause a stagnation in our technological and intellectual progress. Rather than striving for new solutions and breakthroughs to solve humanity’s problems, there is the risk that we will just rely on machines to maintain the status quo. The terrifying result may even be a halt in humanity’s evolution.
As a species that relies exclusively on the finite resources of our planet and sun to support us, we risk long term extinction if we stop innovating and pushing technological boundaries, both at home and out into space.
So, what can be done to prevent such concerns from materializing? Do we simply ban the continued development of AI as some have suggested? We don’t believe that that is the answer. The first step is to get the balance right between AI assistance and human creativity to ensure that our ability to innovate and evolve is enhanced rather than hindered. This will require clear guidelines for educators and regulators on how we can harness the potential of AI safely and effectively. Students using ChatGPT to cheat on assignments may seem harmless, but it sets a dangerous precedent that we need to address quickly.
While some businesses and institutions may be tempted to cut short-term costs by replacing scientists, innovators, and creatives with AI alternatives, the long-term implications of this move may be disastrous. Legislation may be required to prevent this, but what is really required is widespread understanding about the positive and negative attributes of artificial intelligence and machine learning.
Can we really turn our backs on the hugely positive contribution of AI towards disease identification and other medical advancements? Or the role that it will play in monitoring and mitigating the risks of climate change and resource depletion?
We believe that we shouldn’t have to if the focus of AI development is on its use as a tool to enhance and accelerate the work of scientists, engineers, and other professionals. Getting the balance right could see AI quite literally become the savior of the human race rather than cause its destruction. Hollywood will just need to find a new bogeyman to scare us with.
Customer service automation: Advantages and examples
Explore how customer service automation can empower your support strategy and help your customers get the answers they’re looking for – when and how they want.
How to build a Facebook Messenger chatbot
Everything you need to know about creating a chatbot for Facebook Messenger, including the unique Messenger benefits you can tap into, and step-by-step instructions on how to build your own chatbot with no coding knowledge required.
What is a chatbot template and how templates can help your business
Read on to learn more about chatbot templates: benefits, usage, and best practices for seamless integration.